According to CCTV reports, Ms. Peng's medical insurance account was fraudulently charged more than 40,000 HKD without her knowledge. Preliminary investigations 

According to CCTV reports, Ms. Peng's medical insurance account was fraudulently charged more than 40,000 HKD without her knowledge. Preliminary investigations by the police revealed that a criminal group used advanced AI technology to create a fake video of Ms. Peng's face, successfully bypassing her medical insurance account's face verification and carrying out a series of fraudulent transactions.

![2025011503.png](https://cdn.aisecurius.com/images/fc8/fc8403c8-9c9d-4ce6-8b41-6f77f3dc1171.png)

# Case Overview

The criminals first searched the internet for Ms. Peng's photos to obtain her personal data. Using AI technology, they transformed her photos into a realistic facial video. Then, the group used these AI-generated videos to successfully bypass the facial recognition feature of the medical insurance system. With the fake facial video, the criminals were able to log into Ms. Peng's medical insurance account and begin using her account funds to purchase medications and other medical services.

**Steps Taken by the Criminals:**

1. **Collecting Personal Photos**: The criminals obtained publicly available photos of Ms. Peng via the internet. These photos served as the foundation for generating the fake video.

2. **AI Impersonation of the Face**: Using AI face-swapping technology, the criminals transformed Ms. Peng’s photos into a lifelike facial video. The video looked very natural and was able to deceive the identity verification system.

3. **Illegal Login to the Medical Insurance Account**: With the forged facial video, the criminals successfully bypassed the facial recognition system of the medical insurance platform and accessed Ms. Peng’s account.

4. **Stealing Account Funds**: Once logged in, the criminals started making large purchases at pharmacies and other places using Ms. Peng’s medical insurance card, totaling more than 40,000 HKD in stolen funds.

The police investigation revealed that this case was not an isolated incident. The criminal group had previously used similar methods to fraudulently charge multiple doctors' medical insurance accounts, with nearly 20 cases reported, and a total stolen amount exceeding 500,000 HKD.


# Self-Protection Advice for Users


This case not only exposes the potential risks of AI technology misuse but also reminds both platform operators and users to remain vigilant and adopt effective measures to protect personal information security in an increasingly complex online environment. In response to evolving AI-driven fraud methods, protective measures should address both technological defenses and behavioral habits to ensure that personal information is not easily leaked. Below are some practical self-protection suggestions to help users effectively cope with AI-driven cyber fraud.

1. **Enhance Identity Verification**: Enable two-factor authentication (2FA) or multi-factor authentication (MFA), such as SMS codes, dynamic passwords, or email verification, to increase account security. Even if biometric systems (such as fingerprints or facial recognition) are compromised, additional layers of verification can still prevent unauthorized logins. For payment accounts and social media platforms, users should not only enter passwords but also confirm with a one-time verification code received on their phones to increase security. Even if biometric information is stolen, enabling additional identity verification methods can effectively prevent illegal access.

2. **Limit Sharing of Sensitive Information**: Avoid sharing sensitive information such as front-facing photos or videos on social media or public platforms. Such data can be exploited by AI to fabricate facial features or generate false profiles. For example, avoid uploading too many facial images, especially selfies or full-body photos that could be used for face-swapping technology. Also, when enabling facial or fingerprint recognition, choose highly secure technology providers and avoid using unverified or unreliable tools. Reducing unnecessary exposure of personal information can effectively lower the risk of misuse.

3. **Regularly Check Account Security**: Periodically check login records provided by platforms to look for any unusual devices or locations. If any suspicious activity is found, immediately change your account password and notify the platform for further action. Most banks, payment platforms, and social media services offer login history checks, so users can regularly review their account login history and check for any unknown devices or logins from unusual locations. If any anomalies are found, change your password immediately and enable additional security settings (such as device binding or security question verification).

4. **Increase Awareness of AI Technology**: Learn about common characteristics of AI-generated fake videos and audio. Improving your ability to recognize AI-manipulated content can help you make the right judgment when facing potential scams.

By following these measures, users can effectively reduce the risk of AI fraud and improve their information security through a combination of technology and behavior. By staying vigilant and strengthening security awareness, users can better protect their personal privacy in the digital age and avoid falling victim to fraud.


# How Platforms Can Strengthen Prevention?

Fraudsters are using a combination of technology and psychology to precisely manipulate their victims. In the face of AI-driven scams, society must work together through technological means, education, and emotional support to build a protective wall and prevent more people from falling victim. Short video platforms, in particular, need to adopt multiple technical measures to identify fraudulent accounts at the source.

Identify Abnormal Devices
Dingxiang Device Fingerprinting records and compares devices to distinguish between legitimate users and potential fraudulent behavior. This technology uniquely identifies and recognizes each device, identifying maliciously controlled devices such as virtual machines, proxy servers, and simulators. It analyzes whether the device shows abnormal or inconsistent behaviors, such as logging in with multiple accounts, frequently changing IP addresses, or altering device attributes, helping to track and identify fraudulent activities.

Identify Accounts with Abnormal Operations
Continuous identity verification during the session is crucial to ensure the user’s identity remains consistent throughout the usage. Dingxiang atbCAPTCHA quickly and accurately distinguishes whether the operator is a human or a machine, precisely identifying fraudulent activities, monitoring in real time, and intercepting abnormal behavior.

Prevent Fake Videos from Face-Swapping
Dingxiang's full-link panoramic face security threat perception solution conducts intelligent verification across multiple dimensions, including device environment, facial information, image authentication, user behavior, and interaction status. It quickly identifies over 30 types of malicious attacks, such as injection attacks, live-body forgery, image forgery, camera hijacking, debugging risks, memory tampering, Root/Jailbreak, malicious ROM, simulators, etc. Once fake videos, fraudulent facial images, or abnormal interactions are detected, the system automatically blocks the operation. It also allows flexible configuration of video verification strength and user-friendliness, enabling a dynamic mechanism that strengthens verification for abnormal users while maintaining regular atbCAPTCHA for normal users.

Identify Potential Fraud Threats
Dingxiang Dinsight helps enterprises with risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. The average processing speed of Dinsight’s daily risk control strategies is under 100 milliseconds, supporting configurable multi-party data access and accumulation. Based on mature indicators, strategies, models, and deep learning technologies, it enables self-monitoring and self-iteration of risk control performance. Dinsight, paired with the Xintell smart model platform, optimizes security strategies for known risks, analyzes potential risks through data mining, and configures risk control strategies for different scenarios with one click. Using association networks and deep learning technologies, the platform standardizes complex data processing, mining, and machine learning processes, offering a one-stop modeling service from data processing and feature derivation to model construction and deployment.


AI Impersonates a Doctor’s Face to Steal 40,000 HKD from Medical Insurance Account

 His short videos and live-streamed lectures have become widely circulated, effectively raising the public's awareness about fraud, making him a true “internet celebrity cop” in the show.

In the popular TV show *The Trident 2* on Dragon TV, the character played by actor Chen Jianbin, Cui Tiejun, is a seasoned, upright police officer. Meanwhile, he also embraces the times and fully utilizes new media platforms, becoming an “anti-fraud internet celebrity.” Cui Tiejun uses sharp language, professional knowledge, and vivid case studies to conduct anti-fraud advocacy online, revealing various scam tactics and reminding the public to stay vigilant. His short videos and live-streamed lectures have become widely circulated, effectively raising the public's awareness about fraud, making him a true “internet celebrity cop” in the show.

![2025020701.jpg](https://cdn.aisecurius.com/images/98b/98bea0f4-fa4d-4b17-b09c-39dd9a1df59a.jpg)

As the saying goes, "fame brings controversy." Cui Tiejun’s growing online influence soon caught the attention of criminals. In *The Trident 2*, the Anti-Fraud Center suddenly receives a report from the public. The reporter claimed they had received a phone call from someone pretending to be “Cui Tiejun.” The voice on the phone sounded exactly like the anti-fraud internet celebrity. The impersonated “Cui Tiejun” used various reasons to persuade the victim to make a transfer. At first, the victim was suspicious, but when they heard the familiar voice of “Cui Tiejun” and his professional explanation of anti-fraud knowledge, they gradually relaxed their guard and eventually transferred their hard-earned savings to the account provided by the scammer.

Furious, Cui Tiejun decided to call the scammer himself to confront the imposter. Upon answering the call, he was shocked and enraged to hear a voice on the other end that was exactly like his own! The scammer had used AI technology to perfectly clone his voice, even mimicking his speaking habits with incredible precision.

What was even more infuriating was that the fake “Cui Tiejun” on the other end showed no remorse and even taunted the real Cui Tiejun for his “internet celebrity” status. The imposter even boasted about using his fame to carry out more fraud. Faced with such bold provocation, Cui Tiejun was filled with rage but also felt a deep sense of helplessness. He realized that the criminals behind this were not just ordinary fraudsters, but highly sophisticated criminals armed with advanced technology, making them more cunning and difficult to deal with.

This “AI Impersonation Scam” story in *The Trident 2* is not just a simple case; it reflects the harsh reality that new forms of criminal tactics are constantly emerging in the age of rapid technological advancement. While AI technology brings convenience to society, it is also exploited by criminals as a tool for crime. AI voice cloning and deepfake technologies provide new “weapons” for fraudsters, making their tactics more covert and harder to defend against.


# How to Effectively Prevent Impersonation Scams by Celebrities?


In these new forms of telecom fraud, scammers use AI-generated realistic audio and video, making it extremely difficult for victims to identify abnormalities through their intuition. Therefore, even when verifying the identity of the person over the phone or through video, it is crucial to take multiple steps to confirm the identity, especially when it involves financial transactions, to avoid making hasty decisions and falling for the scam. Dingxiang Defense Cloud Business Security Intelligence Center recommends that the public increase vigilance and adopt multi-layered verification measures.

1. **Call Back to Verify Identity**: Hang up the incoming call and contact the person again using a reliable method you already know (e.g., a phone number saved in your contact list or a contact from the official website). Do not contact the number provided by the caller again, as the scammer may use a spoofed caller ID.

2. **Consult Friends and Family**: If you receive a phone call or message asking you to donate or transfer money, check with your friends and family to see if they have received similar requests or if such an event has occurred. If you are unsure, ask them to help verify the situation.

3. **Increase Awareness of AI Forged Technologies**: When receiving a call, even if the voice sounds familiar, if there are inconsistencies in tone, speech fluency, or emotional expression, you should immediately be suspicious. AI-generated voices may lack certain details, such as tone variations and emotional authenticity.

4. **Report to Authorities**: Dial the local emergency number and report the fraud in detail to the police.

These preventive measures can help individuals resist the threat of AI fraud and remind everyone to stay calm and vigilant when facing urgent financial requests, avoiding falling into the trap set by criminals.

Scammers Use AI to Impersonate "Chen Jianbin" and Trick the Actor Himself

The police investigation found that after obtaining Ms. Li's personal medical insurance information, the criminals used AI technology to forge Ms. Li's face, successfully logged into Ms. 

In December 2024, Ms. Li, from Kaiyang County, Guiyang City, encountered an unbelievable online fraud incident - her medical insurance account suddenly received a deduction notification, showing that she had purchased over 10,000 yuan worth of precious Chinese medicinal materials in another place. The investigation found that this was a case of criminals using AI face-swapping technology to steal from personal medical insurance accounts. The police quickly tracked down the criminal gang and successfully apprehended them.


# Ms. Li's Medical Insurance Card Was Stolen and Swiped


The police investigation found that after obtaining Ms. Li's personal medical insurance information, the criminals used AI technology to forge Ms. Li's face, successfully logged into Ms. Li's medical insurance card, and then carried out a frenzied theft.

![2025020601.png](https://cdn.aisecurius.com/images/8e9/8e976ddf-ee3b-4334-862b-9ae83feae399.png)

**The Steps Involved:**

1. **Obtain Medical Insurance Information:** The criminals obtained Ms. Li's medical insurance account information through illegal channels. This information may have been obtained through social media, data breaches, phishing websites, false telecommunications fraud, etc.

2. **Forge Face through AI Face-Swapping:** The criminals used AI face-swapping technology to successfully forge a video of Ms. Li's face.

3. **Log in to the Medical Insurance APP in a Different Place:** The criminals used the forged face video to successfully log into Ms. Li's medical insurance account.

4. **Steal from the Medical Insurance Account:** The criminals obtained Ms. Li's electronic medical insurance voucher (QR code) and purchased over 10,000 yuan worth of precious Chinese medicinal materials, including "Dong'e Donkey-hide Gelatin" and "Pianzaihuang" and other high-priced drugs, at a pharmacy.


# Self-Protection Suggestions for Users


This case not only exposed the potential risks of the abuse of AI technology but also reminded platform operators and users to be vigilant in the increasingly complex network environment and take effective measures to protect personal information security and resist the network security threats brought by new technologies. To deal with the constantly evolving AI fraud methods, preventive measures should be taken from both technical protection and behavioral habits to ensure that personal information is not easily leaked. Here are some practical self-protection suggestions to help users effectively deal with AI-driven cyber fraud.

1. **Strengthen Identity Verification:** Enable two-step verification (2FA) or multi-factor authentication (MFA), such as SMS verification codes, dynamic passwords, email verification, etc., to increase the security of your account. Even if the biometric system (such as fingerprint or facial recognition) is breached, additional verification layers can still effectively prevent unauthorized logins. For example, for payment accounts, social media platforms, etc., when logging in, users need to enter a one-time verification code received by their mobile phone in addition to their password for confirmation, thereby increasing security. Even if the account's biometric information is stolen, enabling additional authentication methods can still effectively prevent the account from being illegally logged in.

2. **Reduce Sharing of Sensitive Information:** Try to avoid sharing sensitive information such as photos or videos with frontal facial photos in public places such as social media and online platforms. This information may be used by AI technology to forge personal facial features or generate false information. For example, do not upload too many photos with facial features on open social platforms, especially selfies, full-body photos, etc. that may be used by AI for face-swapping technology. At the same time, when enabling facial recognition or fingerprint verification, choose those with highly secure technology providers, and avoid using unverified or untrustworthy tools. Reducing unnecessary exposure of personal information can effectively reduce the risk of being abused.

3. **Regularly Check Account Security:** Regularly check the login records provided by the platform to check for abnormal devices or location logins. Once any abnormal behavior is found, immediately change the account password and promptly notify the platform to take further measures. Most banks, payment platforms, and social media provide login record query functions, and users can regularly check their account's login history to check for unknown devices or logins from different places. If any abnormalities are found, change the password immediately and turn on more security settings (such as device binding or security question verification).

4. **Improve the Ability to Identify AI Technology:** Understanding some common features of AI-generated fake videos and audio, and improving the ability to identify AI-forged content can help you make the right judgments when faced with potential scams.

Through the above measures, users can effectively reduce the risk of encountering AI fraud, and improve their information security protection capabilities through the combination of technology and behavior. Only by staying vigilant and strengthening security awareness can we effectively protect personal privacy in the digital age and avoid becoming victims of fraud.


# How Platforms Can Strengthen Prevention?

Fraudsters are using a combination of technology and psychology to precisely manipulate their victims. In the face of AI-driven scams, society must work together through technological means, education, and emotional support to build a protective wall and prevent more people from falling victim. Short video platforms, in particular, need to adopt multiple technical measures to identify fraudulent accounts at the source.

Identify Abnormal Devices
Dingxiang Device Fingerprinting records and compares devices to distinguish between legitimate users and potential fraudulent behavior. This technology uniquely identifies and recognizes each device, identifying maliciously controlled devices such as virtual machines, proxy servers, and simulators. It analyzes whether the device shows abnormal or inconsistent behaviors, such as logging in with multiple accounts, frequently changing IP addresses, or altering device attributes, helping to track and identify fraudulent activities.

Identify Accounts with Abnormal Operations
Continuous identity verification during the session is crucial to ensure the user’s identity remains consistent throughout the usage. Dingxiang atbCAPTCHA quickly and accurately distinguishes whether the operator is a human or a machine, precisely identifying fraudulent activities, monitoring in real time, and intercepting abnormal behavior.

Prevent Fake Videos from Face-Swapping
Dingxiang's full-link panoramic face security threat perception solution conducts intelligent verification across multiple dimensions, including device environment, facial information, image authentication, user behavior, and interaction status. It quickly identifies over 30 types of malicious attacks, such as injection attacks, live-body forgery, image forgery, camera hijacking, debugging risks, memory tampering, Root/Jailbreak, malicious ROM, simulators, etc. Once fake videos, fraudulent facial images, or abnormal interactions are detected, the system automatically blocks the operation. It also allows flexible configuration of video verification strength and user-friendliness, enabling a dynamic mechanism that strengthens verification for abnormal users while maintaining regular atbCAPTCHA for normal users.

Identify Potential Fraud Threats
Dingxiang Dinsight helps enterprises with risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. The average processing speed of Dinsight’s daily risk control strategies is under 100 milliseconds, supporting configurable multi-party data access and accumulation. Based on mature indicators, strategies, models, and deep learning technologies, it enables self-monitoring and self-iteration of risk control performance. Dinsight, paired with the Xintell smart model platform, optimizes security strategies for known risks, analyzes potential risks through data mining, and configures risk control strategies for different scenarios with one click. Using association networks and deep learning technologies, the platform standardizes complex data processing, mining, and machine learning processes, offering a one-stop modeling service from data processing and feature derivation to model construction and deployment.


Scammers Use AI to Forge Face-Swap Videos, Log in to Others' Medical Insurance Cards

Through its collaboration with Dingxiang, the Zhangxin Changxing App has not only enhanced its security, but also mitigated potential security risks,

The Zhangxin Changxing App is a comprehensive multifunctional mobile app that integrates local news, convenience information, public services, government services, local services, video playback, and live streaming functions. It provides users with the latest news, convenience services, and access to video live streaming. Compliance has become a key factor in the development of the app. For app developers, compliance means strengthening the protection of user data, implementing appropriate security measures and technologies, and ensuring user privacy. Additionally, it is important to adhere to relevant laws and regulations, clearly define user permissions, and update privacy terms and user agreements in a timely manner, providing users with sufficient transparency and choice.

![2025020401.png](https://cdn.aisecurius.com/images/a7c/a7cc614a-8a2a-45d4-bb32-6a720da420d5.png)

To further enhance the app's security and compliance, the Zhangxin Changxing App has deployed Dingxiang App Hardening. Dingxiang App Hardening is an adaptive hardening solution based on graph neural network technology. It conducts in-depth analysis and feature extraction of app code, automatically selecting the appropriate methods for encryption, obfuscation, and compression. This increases the difficulty of reverse engineering the code, effectively preventing the app from being cracked, copied, or repackaged by attackers.

With its unique code obfuscation technology, Dingxiang App Hardening ensures that the app’s source code is harder to crack and reverse-engineer, protecting against threats from hackers and malicious software, and reducing security risks arising from system vulnerabilities. It further protects sensitive data within the app, including users' personal and transaction data, preventing data leakage or misuse. Additionally, Dingxiang App Hardening monitors terminal devices, operating environments, and user behavior to detect anomalies and issue warnings in real-time, blocking potential security threats.

Through its collaboration with Dingxiang, the Zhangxin Changxing App has not only enhanced its security, but also mitigated potential security risks, allowing it to provide users with a better service experience within a legally compliant framework. It offers users a safer, more reliable, and compliant service, supporting the comprehensive development of the digital era.

Dingxiang Assists in the Security and Compliance of the Zhangxin Changxing App

In November 2024, an 18-year-old mainland female student studying at the University of Hong Kong received a call from someone claiming to be a "mainland law enforcement officer," accusing her 

In November 2024, an 18-year-old mainland female student studying at the University of Hong Kong received a call from someone claiming to be a "mainland law enforcement officer," accusing her of being involved in a money laundering case and demanding she transfer approximately HKD 9.2 million to multiple bank accounts for investigation. Later, the student realized she had been scammed after she could not contact the caller.

![2025012402.jpg](https://cdn.aisecurius.com/images/43f/43f3633e-d5ad-493a-91b3-7ed72105d0fc.jpg)

The student, surnamed Hong, came to Hong Kong from mainland China in September 2024. On November 27 of the same year, she received a call from someone claiming to be a telecommunications company employee, accusing her of being involved in a money laundering case, and the call was then transferred to someone claiming to be a "mainland law enforcement officer." The fraudster instructed her to download an app for monitoring, and from December 2, 2024, to January 4, 2025, she transferred a total of approximately HKD 9.2 million to 3-5 local accounts provided by the fraudster, claiming it was for investigation. After the fraudster lost contact with her, she realized it was a scam and reported it to the police on the 9th of the same month. The Criminal Investigation Team of the Hong Kong Western District is investigating the case and has issued arrest warrants for five suspects.

This incident is not an isolated case. Emails sent to students by the University of Hong Kong show that more than 60 students have become victims of fraud in recent months, with a total loss of over HKD 60 million. The majority of the victims were first-year mainland students, and the scams occurred mainly between September and October last year, when they were unfamiliar with Hong Kong's environment and culture, making them easy targets. According to police statistics, from January to November 2024, Hong Kong reported 40,470 fraud cases, an increase of 8.7% year-on-year, with a total amount of HKD 8.52 billion involved. Among them, phone fraud cases increased by 172.8%, reaching 5,168 cases. From September 2024 to November 2024, the Hong Kong Island district recorded 55 cases of "imposter official" scams involving "Mainland students," with total losses amounting to HKD 39 million.


# Analysis of Reasons Why 18-Year-Old Students Are Easily Scammed


There are multiple reasons why mainland university students are vulnerable to scams. These include personal experiences, psychological vulnerability, cultural differences, inadequate anti-fraud education, and the increasing complexity of scam tactics.

![2025012401.png](https://cdn.aisecurius.com/images/b72/b72902e6-37d3-429f-b619-8bf8b6655e22.png)

1. **Unfamiliarity with the local environment.** For many mainland students studying in Hong Kong, it is a completely new environment for both living and studying. The unfamiliarity with local culture, legal systems, and social norms increases their vulnerability when facing external threats. When fraudsters impersonate "law enforcement officers," it becomes easier for students to trust them and follow their instructions.

2. **Lack of social experience.** Many mainland students, although academically successful, lack independent living experience and the ability to deal with complex situations. Having not encountered real-life anti-fraud information or the various scams, they find it difficult to remain calm when receiving unexpected calls, making them more susceptible to scammers' words. This lack of social experience makes them easy targets for scammers, who can easily manipulate them.

3. **High-fidelity scam techniques.** Fraudsters not only impersonate "law enforcement officers" using phone voices but also use the credibility of institutions like "law" and "banks" to create fear in the victims. Scammers often force victims to react quickly by creating an "urgent situation," manipulating the victims' emotions, making it hard for them to think rationally. Victims believe the fraudsters' "official" identities without suspicion, and in the face of the pressure and urgency, they fail to verify the information and are scammed.

4. **Lack of anti-fraud education and information.** Although the University of Hong Kong has provided some anti-fraud publicity, many victims have not been exposed to relevant anti-fraud knowledge. Especially at the beginning of the school year, anti-fraud education was insufficient. Mainland students often focus on their academic performance and course selection rather than paying attention to anti-fraud knowledge and life skills. Given the increasing variety of scams, simple verbal reminders and sporadic publicity are clearly insufficient to capture students' attention. Many new students lack awareness of security issues related to social environments and networks, making them more vulnerable to scams.

To address these problems, society and educational institutions need to pay more attention to cultivating students' security awareness, help students improve their ability to identify fraud, and provide adequate safety education.


# How to Effectively Prevent Phone Scams?


In this new type of online telecom fraud, scammers use AI-generated realistic audio and video, making it difficult for victims to detect abnormalities through intuition. Therefore, even when verifying the identity of the other party through phone or video, it is crucial to confirm their identity using multiple methods to avoid making rushed decisions and falling for scams. Dingxiang Defense Cloud Business Security Intelligence Center advises the public to stay alert and adopt multi-layered verification measures.

![2025012403.jpg](https://cdn.aisecurius.com/images/d40/d401e70f-bc2d-447c-aae2-4edb3c9f5250.jpg)

1. **Call back to verify identity.** Hang up the incoming call and contact the other party through known and reliable contact information (such as phone numbers stored in your phone or official contact information provided on official websites). Do not contact them again using the number they provide, as scammers may use fake caller ID numbers.

2. **Consult teachers, classmates, or friends.** If you receive a call or message asking you to donate money or transfer funds, it is advisable to check with friends and family to see if they received similar requests or if similar incidents have occurred. If you cannot determine this yourself, ask for help in verifying.

3. **Increase awareness of AI forgery techniques.** When receiving a call, even if the voice sounds familiar, if the tone or fluency of language, emotional expression, etc., are inconsistent, you should immediately be suspicious. AI-generated voices may lack details such as tone fluctuations or emotional authenticity.

4. **Call the police for help.** Dial the local emergency number to report the fraud to the police in detail.

These preventive measures will not only help individuals defend against AI fraud threats but also remind everyone to remain calm and vigilant when faced with sudden financial requests, avoiding falling into the trap of criminals.

**Platforms Should Strengthen Security Warnings**

Fraudsters are using a combination of technology and psychology to manipulate victims. In the face of AI-generated fraud, society must work together to build strong defenses through technology, education, and emotional support to prevent more people from being deceived. In particular, short video platforms need to adopt multiple technologies and methods to identify fraudster accounts and effectively detect fake celebrity videos.

1. **Identify abnormal devices.** Dingxiang Device Fingerprinting can distinguish between legitimate users and potential fraud by recording and comparing device fingerprints. This technology uniquely identifies each device and detects maliciously controlled devices such as virtual machines, proxy servers, and simulators. It analyzes whether the device has unusual behaviors, such as multi-account logins, frequent IP address changes, or device attribute changes, helping to track and identify fraudulent activities.

2. **Identify abnormal account activities.** Remote logins, device changes, phone number changes, dormant accounts suddenly becoming active, etc., require frequent verification; in addition, continuous identity verification during a session is crucial to ensure the user's identity remains consistent throughout the session. Dingxiang atbCAPTCHA can quickly and accurately differentiate between human and machine operators, precisely identifying fraudulent behaviors and monitoring abnormal activities in real time.

3. **Prevent fake face-swapping videos.** Dingxiang’s full-link panoramic face security threat perception solution intelligently verifies multiple dimensions of information, including device environment, facial data, image authentication, user behavior, and interaction status. It quickly identifies injection attacks, live-face forgery, image forgery, camera hijacking, debugging risks, memory tampering, Root/Jailbreak, malicious ROM, simulators, and over 30 types of malicious attacks. Once fake videos or fraudulent faces are detected, it can automatically block the operation. Additionally, it can dynamically adjust video verification strength and user-friendly settings, strengthening verification for abnormal users and implementing a dynamic mechanism.

4. **Dig into potential fraud threats.** Dingxiang Dinsight helps enterprises with risk assessments, anti-fraud analysis, and real-time monitoring to improve efficiency and accuracy in risk control. The average processing time of Dinsight’s daily risk control strategies is under 100 milliseconds. It supports the integration of multiple data sources and can automatically optimize security strategies for known risks. Based on risk logs and data mining, it can identify potential risks, configure different scenarios, and support risk control strategies.
5. 
# How Platforms Can Strengthen Prevention?


Fraudsters are using a combination of technology and psychology to precisely manipulate their victims. In the face of AI-driven scams, society must work together through technological means, education, and emotional support to build a protective wall and prevent more people from falling victim. Short video platforms, in particular, need to adopt multiple technical measures to identify fraudulent accounts at the source.

1. **Identify Abnormal Devices**  
Dingxiang Device Fingerprinting records and compares devices to distinguish between legitimate users and potential fraudulent behavior. This technology uniquely identifies and recognizes each device, identifying maliciously controlled devices such as virtual machines, proxy servers, and simulators. It analyzes whether the device shows abnormal or inconsistent behaviors, such as logging in with multiple accounts, frequently changing IP addresses, or altering device attributes, helping to track and identify fraudulent activities.

2. **Identify Accounts with Abnormal Operations**  
Continuous identity verification during the session is crucial to ensure the user’s identity remains consistent throughout the usage. Dingxiang atbCAPTCHA quickly and accurately distinguishes whether the operator is a human or a machine, precisely identifying fraudulent activities, monitoring in real time, and intercepting abnormal behavior.

3. **Prevent Fake Videos from Face-Swapping**  
Dingxiang's full-link panoramic face security threat perception solution conducts intelligent verification across multiple dimensions, including device environment, facial information, image authentication, user behavior, and interaction status. It quickly identifies over 30 types of malicious attacks, such as injection attacks, live-body forgery, image forgery, camera hijacking, debugging risks, memory tampering, Root/Jailbreak, malicious ROM, simulators, etc. Once fake videos, fraudulent facial images, or abnormal interactions are detected, the system automatically blocks the operation. It also allows flexible configuration of video verification strength and user-friendliness, enabling a dynamic mechanism that strengthens verification for abnormal users while maintaining regular atbCAPTCHA for normal users.

4. **Identify Potential Fraud Threats**  
Dingxiang Dinsight helps enterprises with risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. The average processing speed of Dinsight’s daily risk control strategies is under 100 milliseconds, supporting configurable multi-party data access and accumulation. Based on mature indicators, strategies, models, and deep learning technologies, it enables self-monitoring and self-iteration of risk control performance. Dinsight, paired with the Xintell smart model platform, optimizes security strategies for known risks, analyzes potential risks through data mining, and configures risk control strategies for different scenarios with one click. Using association networks and deep learning technologies, the platform standardizes complex data processing, mining, and machine learning processes, offering a one-stop modeling service from data processing and feature derivation to model construction and deployment.

18-Year-Old Female College Student Scammed of HKD 9.2 Million: Analysis of the Reasons

Recently, an 82-year-old man, Mr. Wang from Hantai County, Shandong, received a phone call from abroad and was scammed out of 326,000 yuan. 

Recently, an 82-year-old man, Mr. Wang from Hantai County, Shandong, received a phone call from abroad and was scammed out of 326,000 yuan. The fraudster used AI technology to synthesize the voice of Mr. Wang's grandson and impersonated him to ask for tuition fees, ultimately leading Mr. Wang to transfer a large sum of money to the fraudster’s account without verifying the request.


# Using AI Synthesis to Deceive the Victim


The incident occurred a few days ago when Mr. Wang received a phone call from abroad. The voice on the other end sounded just like his grandson, and Mr. Wang immediately recognized the familiar tone. However, the “grandson” on the phone seemed to urgently need funds, claiming that due to tuition issues, the school had threatened to expel him, and if the money was not sent quickly, he would not be able to continue his studies.

![2025012001.png](https://cdn.aisecurius.com/images/196/19690be8-9525-4f2c-9fb0-f40c89ef51ba.png)

To make Mr. Wang believe the situation, the fraudster not only mimicked the grandson's voice but also cleverly incorporated the name of the school his grandson attended, the tuition fee amount, and some family details. These pieces of information convinced Mr. Wang that everything was real, and he did not suspect anything unusual about the caller’s identity.

The fraudster deliberately created a sense of urgency during the conversation, repeatedly stressing that time was running out, and if the money was not sent immediately, the grandson would face serious consequences. Mr. Wang's emotions were stirred, and eager to help his grandson, he transferred 326,000 yuan to the specified account under the fraudster’s instructions.


# Why Did Mr. Wang Fall for the Scam So Easily?


According to the details disclosed in the case, the reasons why Mr. Wang was successfully deceived can be summarized as follows:

1. **Emotional Dependence and Trust**: Mr. Wang had a very close relationship with his grandson, who held an important place in his heart. The fraudster exploited this emotional bond by synthesizing the grandson's voice, which made it difficult for Mr. Wang to detect anything amiss. His emotional reliance on his grandson led him to quickly respond with a desire to help, without questioning the content of the phone call.

2. **Advanced AI Forgery**: The fraudster not only used the grandson's voice but also employed sophisticated AI technology to create a “deep fake” voice that was nearly indistinguishable from the real one, greatly enhancing the credibility of the call. Mr. Wang heard the voice of his grandson and did not suspect anything, thinking it was a genuine request for help.

![2025010902.png](https://cdn.aisecurius.com/images/5cb/5cbf2e8b-b264-455d-88f4-ba9a032d47bb.png)

3. **Precise Social Engineering Tactics**: In addition to mimicking the voice, the fraudster used social engineering techniques to gather detailed information about the family, such as the name of the school the grandson attended abroad, the amount of tuition fees, and even some private family conversations. These specific and accurate details made Mr. Wang trust the caller even more, believing the request was genuine.

4. **Sense of Urgency and Emotional Manipulation**: The fraudster created a huge sense of urgency over the phone, repeatedly stressing that if the money was not sent quickly, the grandson would face expulsion. In the face of the “urgent need” to help his grandson, Mr. Wang’s emotions were strongly manipulated. His frantic thinking in the moment clouded his judgment, leading him to hastily decide to transfer the money.

5. **Lack of Awareness of New Fraud Techniques**: Mr. Wang did not realize the complexity of this new type of online phone scam, especially the use of AI technology in fraud. Having never encountered similar scams before and lacking relevant preventative knowledge and education, he failed to recognize that he was facing a high-tech scam.


# How to Avoid Similar Scams?


This case highlights the high-tech and covert nature of modern scams, especially the use of AI-generated deepfake technology, which makes it difficult for many people, particularly the elderly, to discern the truth. To help elderly individuals effectively prevent such new types of scams, the Dingxiang Defense Cloud Business Security Intelligence Center provides the following practical preventive measures, reminding everyone to stay alert when facing similar urgent requests for money and avoid being manipulated by fraudsters.

1. **Call Back to Confirm the Identity**  
When receiving a suspicious or unfamiliar phone call, the first step is to stay calm and avoid hastily responding to any requests. Especially when faced with urgent requests for money, you can use the "call back" strategy. Hang up under the pretext of unclear signal and avoid making impulsive decisions in a state of panic. Then, call back using known phone numbers of family or friends (e.g., mobile phones, home phones, or other private contact channels) to verify if it’s really a relative or friend asking for help.

2. **Set up Family Authentication to Ensure Real Identity**  
To further ensure identity security, it is recommended that elderly individuals and their family members set up a “safe word” or “challenge question” in advance as a unique method of confirming identity. This way, when receiving a suspicious call, they can ask the safe word or question to determine if the caller is indeed a relative or friend. If the caller cannot answer correctly or tries to avoid the question, the call can be immediately identified as a scam, and the phone should be hung up. This simple and effective identity verification method greatly increases the chances of preventing fraud.

3. **Consult Family and Friends, Don't Make Decisions Alone**  
Fraudsters often use urgent and threatening language to create panic in the victim, speeding up their decision-making process. In such cases, elderly individuals should consult other family or friends before making any financial transfer decisions, helping to calmly analyze the situation. This approach can effectively reduce pressure and allow for a more rational decision.

4. **Report to the Authorities Promptly for Help**  
If identity verification fails or it is confirmed that a scam has occurred, the first step should be to dial the local emergency number (110) and report the situation to the police, helping to track the fraudster’s movements. The police usually have more technical means to assist in the investigation. Prompt reporting can reduce losses and help recover the stolen funds. When reporting, try to provide detailed information about the case, including phone numbers, payment information provided by the fraudster, bank account details, etc., which will help the police with the investigation.

With the development of AI technology, scammers' tactics have become increasingly sophisticated, especially in areas such as voice synthesis and video forgery, significantly increasing the success rate of scams. The elderly are particularly vulnerable to these new types of online fraud. Therefore, family members and society as a whole should strengthen anti-fraud education for elderly individuals, helping them raise awareness and avoid making impulsive decisions in sudden emergencies, which could lead to irreparable losses.


# How Platforms Can Strengthen Prevention?


Fraudsters are using a combination of technology and psychology to precisely manipulate their victims. In the face of AI-driven scams, society must work together through technological means, education, and emotional support to build a protective wall and prevent more people from falling victim. Short video platforms, in particular, need to adopt multiple technical measures to identify fraudulent accounts at the source.

1. **Identify Abnormal Devices**  
Dingxiang Device Fingerprinting records and compares devices to distinguish between legitimate users and potential fraudulent behavior. This technology uniquely identifies and recognizes each device, identifying maliciously controlled devices such as virtual machines, proxy servers, and simulators. It analyzes whether the device shows abnormal or inconsistent behaviors, such as logging in with multiple accounts, frequently changing IP addresses, or altering device attributes, helping to track and identify fraudulent activities.

2. **Identify Accounts with Abnormal Operations**  
Continuous identity verification during the session is crucial to ensure the user’s identity remains consistent throughout the usage. Dingxiang atbCAPTCHA quickly and accurately distinguishes whether the operator is a human or a machine, precisely identifying fraudulent activities, monitoring in real time, and intercepting abnormal behavior.

3. **Prevent Fake Videos from Face-Swapping**  
Dingxiang's full-link panoramic face security threat perception solution conducts intelligent verification across multiple dimensions, including device environment, facial information, image authentication, user behavior, and interaction status. It quickly identifies over 30 types of malicious attacks, such as injection attacks, live-body forgery, image forgery, camera hijacking, debugging risks, memory tampering, Root/Jailbreak, malicious ROM, simulators, etc. Once fake videos, fraudulent facial images, or abnormal interactions are detected, the system automatically blocks the operation. It also allows flexible configuration of video verification strength and user-friendliness, enabling a dynamic mechanism that strengthens verification for abnormal users while maintaining regular atbCAPTCHA for normal users.

4. **Identify Potential Fraud Threats**  
Dingxiang Dinsight helps enterprises with risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. The average processing speed of Dinsight’s daily risk control strategies is under 100 milliseconds, supporting configurable multi-party data access and accumulation. Based on mature indicators, strategies, models, and deep learning technologies, it enables self-monitoring and self-iteration of risk control performance. Dinsight, paired with the Xintell smart model platform, optimizes security strategies for known risks, analyzes potential risks through data mining, and configures risk control strategies for different scenarios with one click. Using association networks and deep learning technologies, the platform standardizes complex data processing, mining, and machine learning processes, offering a one-stop modeling service from data processing and feature derivation to model construction and deployment.

AI Impersonates Grandson, 82-Year-Old Man Scammed Out of 326,000 Yuan

The voices and appearances of political figures are highly public and part of their daily life. As a result, their voices are often recorded and widely disseminated.

On the 15th, Thai Prime Minister Petongtarn revealed to the media that she had also fallen victim to a scam where the perpetrators used AI technology to simulate the voice of a leader from an ASEAN country. The scammers, posing as representatives of a "nation," tried to deceive her into making a donation, and she almost fell for the scam. She shared this incident to raise public awareness about telecommunication fraud.

![2025011602.png](https://cdn.aisecurius.com/images/646/646e2794-90ce-45c4-b2f5-c187e4508484.png)

In her media interview, Petongtarn recalled, "The voice was so similar and very clear; I had heard this voice before." The scammers then asked her to donate, claiming that "Thailand is the only ASEAN country that hasn't donated." This raised her suspicions. It wasn’t until they sent a bank account that didn’t belong to the country they claimed to represent that she realized it was a scam.

Petongtarn has instructed the Deputy Prime Minister and Minister of Digital Economy and Society, Phatthana, to handle the case and reminded the public to stay vigilant about telecom fraud.


#  How to Protect Against Impersonation Scams Using AI?


The voices and appearances of political figures are highly public and part of their daily life. As a result, their voices are often recorded and widely disseminated. These audio recordings are easily spread, giving scammers the opportunity to gather enough audio material for synthetic voice generation.

![2024090401.jpeg](https://cdn.aisecurius.com/images/b3d/b3d8baf3-0f9f-4c53-9667-17d4ad081f59.jpeg)

To protect against AI impersonation scams involving political figures or celebrities, Dingxiang’s Defense Cloud Business Security Intelligence Center advises the public to remain alert and adopt multi-layered verification measures.

1. **Call Back to Verify Identity**  
Hang up the phone and reconnect using a known, reliable contact method (e.g., a family member’s number saved in your phone or a contact provided on an official website). Do not call back using the number provided by the caller, as scammers may use spoofed caller IDs.

2. **Contact the Person or Their Team Through Other Means**  
If you receive a call claiming to be from a political figure or celebrity, try contacting the relevant agency or team directly to confirm the authenticity of the request, preventing a hasty decision under the influence of scammers.

3. **Consult Trusted Family or Friends**  
If you receive a call or message asking you to donate or transfer money, consult trusted family or friends to see if they’ve received similar requests or if there are any indications of similar events. If you're unsure, ask for their help in verifying the situation.

4. **Increase Awareness of AI Forgery Technology**  
When receiving a call, even if the voice seems familiar, pay attention to inconsistencies in tone, fluency, or emotional expression. AI-generated voices may lack some details, such as variations in tone or emotional authenticity. If you notice any inconsistencies, remain skeptical and cautious.

Thai Prime Minister Reveals Her Near Fall for AI Scam

Scammers use AI face-swapping technology to replace their faces with those of relatives and even synthesize their voices.

Scammers use AI face-swapping technology to replace their faces with those of relatives and even synthesize their voices. By leveraging deepfake videos and audio, they accurately mimic the tone and expression of loved ones, easily gaining the victim’s trust and committing fraud. Similarly, impersonating law enforcement officers is a common scam tactic. Scammers use AI technology to fake identities, wear police uniforms, and pose as law enforcement personnel conducting "case investigations," creating a false sense of urgency and coercing victims into compliance.

Even more alarmingly, scammers also impersonate celebrities, especially well-known entrepreneurs and public figures, using video calls to interact with victims, building false trust to carry out financial scams.

# 
 Using AI to Impersonate Celebrities


**Case 1: AI Face-Swapping Scam in Fengtai District, Beijing**  
[In November 2023, a community in Fangzhuang Street, Fengtai District](https://www.aisecurius.com/blog/article/128), Beijing, successfully intercepted a scam involving AI face-swapping technology. Criminals cleverly impersonated Jack Ma, attempting to persuade an elderly person to purchase health products.  

![2025011301.png](https://cdn.aisecurius.com/images/954/9544ee9f-7b09-4434-ac9d-f87a46907e11.png)

The scam began with a phone call to the elderly person, during which the scammers, posing as a "disciple of Jack Ma," promoted a poverty alleviation project and promised high investment returns. To enhance credibility, the scammers suggested a video call with "Jack Ma" and used AI face-swapping technology to disguise themselves as him. Due to the high level of realism, the elderly victim did not suspect anything and genuinely believed they were conversing with Jack Ma. The scammers then asked the victim to download QQ software to proceed with the investment.  

Fortunately, community staff and police detected the scam in time. They arranged another video call with the elderly person, exposing the scammer’s true identity. The scammer panicked and quickly ended the call when police appeared, preventing the elderly person from suffering financial losses. The case was resolved promptly.  

**Case 2: AI-Synthesized Celebrity Scam in Yichun, Jiangxi**  
The Public Security Bureau of Yichun City,[ Jiangxi Province, received a call for help from a bank reporting that an elderly person insisted on taking out a loan and refused to leave without doing so](https://www.aisecurius.com/blog/article/125). Upon questioning, the police learned that the elderly person wanted to borrow 2 million yuan to support her so-called “boyfriend,” Jin Dong, in producing a film.  
![2025011302.png](https://cdn.aisecurius.com/images/543/543ea65d-f69f-425e-b9d0-9b1014e61821.png)
The elderly victim claimed she had met “Jin Dong” on a short video platform and developed a romantic relationship with him. The scammer even posted photos and videos together, claiming that the funds were necessary for filming.  

After careful investigation, the police discovered that “Jin Dong” was entirely fabricated, with the photos and videos being AI-generated. Despite this, the elderly victim remained convinced she was in a relationship with the celebrity. Police patiently explained the principles of AI synthesis and demonstrated how the technology worked, ultimately making the elderly person realize she was being scammed and deciding to abandon the loan application.  

These cases highlight the use of AI face-swapping and synthesis technology in scams, especially in emotionally manipulating the elderly. Scammers, disguised as celebrities, exploit the elderly's trust in family and idols to lure them into fraudulent activities.  

# Using AI to Impersonate Law Enforcement


**Case 1: AI-Impersonated Police Scam in Wuxi**  
In May 2024, 75-year-old Ms. Hu received a call from someone claiming to be a police officer from the Wuxi Public Security [Bureau in Jiangsu Province. The scammer alleged that her bank card was involved in a fraud case and requested a financial review](https://www.aisecurius.com/blog/article/126).  
![2025011303.png](https://cdn.aisecurius.com/images/9ea/9ea3a029-90d9-42b9-b2b2-e8812e5c425e.png)
Following the scammer’s instructions, Ms. Hu installed software and joined an online meeting. Unbeknownst to her, she had shared her screen with the scammer, who remotely controlled her phone. Throughout the process, Ms. Hu remained unaware until she checked her account and discovered her funds had disappeared. She then realized she had been scammed and reported the incident to the police.  

Scammers cleverly avoided traditional phone scams by employing remote control and screen-sharing, ensuring the victim was entirely unaware. By posing as law enforcement personnel, the scammers leveraged authority and urgency to gain compliance, leading to financial theft.  

**Case 2: AI Face-Swapping Scam at Hunan Provincial Museum**  
[Recently, the “Lei Feng Volunteer Police” on patrol at the Hunan Provincial Museum noticed an elderly man looking](https://www.aisecurius.com/blog/article/91) distressed while standing outside a mobile service store, engaged in a video call. The volunteer immediately suspected it was a scam and alerted the police.  
![2025011304.png](https://cdn.aisecurius.com/images/7ae/7ae2c334-e3c4-4d0b-bd9a-6ae57db81926.png)

When the police arrived, the elderly man was in a video call with a “police officer.” The scammer, dressed in a police uniform with a stern expression, demanded the man cooperate in a financial review and take notes.  

Upon speaking with the elderly man, the police learned the scammer had contacted him via FaceTime, claiming the elderly man was involved in criminal activities and needed to cooperate with an investigation. The scammer used AI face-swapping technology to replace their face with that of a uniformed “police officer,” increasing credibility. The scammer’s serious tone and fake uniform convinced the man to comply. The police immediately exposed the scam, prompting the scammer to end the call in a hurry.  

These cases underscore how scammers use advanced AI face-swapping and video call technology to convincingly impersonate law enforcement officers. By creating highly realistic videos and disguising their identities, scammers exploit the elderly’s fear of legal authority, manipulating their actions and causing financial losses. Fortunately, timely community intervention and police actions prevented more significant damages.  


#  Why Are the Elderly Particularly Prone to Falling Victim to Scams?


The elderly are especially vulnerable to scams due to a combination of factors, including a lack of understanding about AI technology, trust in authority, emotional dependence, and anxiety-driven decision-making. These factors are outlined below:

1. **Lack of Awareness About AI Technology and Its Risks**  
Many elderly people are unaware of the existence of artificial intelligence technologies, let alone how they are misused in daily life. The advent of AI face-swapping, voice synthesis, and deepfake technology has significantly enhanced the believability of fake information, making it particularly difficult for older individuals to discern between real and fake. Due to their lack of technological literacy, many elderly individuals fail to recognize the risks and potential fraud associated with these technologies. They often assume that videos, voices, or images are genuine, leaving them highly susceptible to scammers' deception.

2. **Fear and Suppression of Rational Judgment**  
Elderly individuals typically have strong traditional values and a high level of trust in authority. Scammers often disguise themselves as public security officers, prosecutors, or court officials, creating a sense of fear and urgency. Upon hearing accusations of alleged legal violations, the elderly are often too afraid to question the scammer’s identity and are easily overwhelmed by the authoritative language used.  
During such moments, fear dominates their decision-making, leading them to comply with the scammer's instructions, which often results in financial losses. For those eager to resolve perceived issues or avoid trouble, scammers exploit their anxiety to create a false sense of urgency, forcing them to make poor decisions under pressure.  

3. **Social Isolation and Emotional Dependence**  
As individuals age, their social circles often shrink, leaving many elderly people with limited companionship. This social isolation makes them emotionally vulnerable and more likely to trust people who feign concern or affection. Scammers impersonate family members, particularly children or grandchildren, to exploit this emotional reliance.  
Additionally, by mimicking well-known public figures such as celebrities or business leaders, scammers enhance the credibility of their schemes. Since celebrities are often associated with trustworthiness and familiarity, elderly individuals may believe the impersonated identity without question, making them prime targets for fraudulent schemes.

4. **Emotional Decision-Making and Reduced Cognitive Agility**  
With age, elderly individuals often experience declines in cognitive function and reaction speed, making them more prone to emotional decision-making during unexpected situations. When faced with unexpected calls or video messages, they may react hastily due to anxiety, failing to verify information thoroughly. In such situations, they may not have the time or capacity to recognize they are being manipulated by scammers.  


#  How Can the Elderly Protect Themselves from AI Scams?  


In these types of scams, which leverage highly realistic AI-generated media, victims often struggle to detect fraud visually or audibly. Therefore, even if video or phone calls are used to "verify" a person's identity, elderly individuals must remain vigilant when confronted with urgent financial requests. Verifying the identity of the person requesting funds through multiple channels is essential to avoid hasty decisions. The Dingxiang Business Security Intelligence Center recommends that individuals enhance their awareness and adopt multi-layered verification measures.  

1. **Call Back to Verify Identity**  
Hang up and reconnect using a known, trusted number (such as one saved in your phone for a family member or a number obtained from an official website). Avoid using any phone numbers provided by the caller, as scammers often use spoofed numbers.

2. **Consult Trusted Friends or Family**  
If you receive a call or message requesting donations or transfers, consult with close family or friends to verify whether others have received similar requests or if such incidents are plausible. If unsure, ask for help in verifying the situation.

3. **Increase Awareness of AI Forgery Technologies**  
Even if the voice on the other end sounds familiar, pay attention to inconsistencies in tone, language, or emotional expressions. AI-generated voices may lack subtle details, such as natural fluctuations in tone or authentic emotional expressions. Any irregularities should raise suspicion.

4. **Report to Authorities for Assistance**  
Dial your local emergency number (e.g., 110 in China) to report any suspected scams to the police in detail.  

These measures not only help individuals defend against AI-related scams but also serve as a reminder to stay calm and vigilant when dealing with sudden financial requests to avoid falling into fraudsters' traps.  


# Platforms Should Strengthen Security Alerts  


Scammers are combining advanced technologies with psychological manipulation to target victims with precision. In response to AI-related scams, collective societal efforts are required to build a robust defense through technology, education, and emotional support. Short video platforms, in particular, need to adopt multi-pronged measures to identify and mitigate fraudulent activities at the source effectively.  

1. **Identifying Suspicious Devices**  
Using **Dingxiang Device Fingerprinting**, platforms can track and identify potential fraudulent devices. By assigning unique identifiers to each device, this technology can detect maliciously manipulated devices, such as virtual machines, proxy servers, and emulators. It also analyzes irregular behaviors, such as multiple account logins, frequent IP changes, or unusual device attributes, to trace and identify fraudulent activities.

2. **Detecting Suspicious Account Activity**  
Monitor for unusual account behaviors, such as remote logins, device changes, phone number changes, or sudden activity from dormant accounts. Continuous identity verification during sessions is crucial to maintain user authentication. **Dingxiang atbCAPTCHA** can distinguish between human and machine operators with speed and accuracy, identifying and blocking fraudulent activity in real-time.

![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)

3. **Detecting Deepfake Videos**  
The **Dingxiang Full-Chain Facial Security Threat Perception Solution** leverages multi-dimensional data—device environment, facial information, image forgery detection, user behavior, and interaction states—to conduct intelligent verification. It can identify injection attacks, spoofing attempts, image manipulation, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, and more than 30 other malicious behaviors. By identifying fake videos, fabricated facial images, and abnormal interactions, the system can automatically block fraudulent activities. It also supports flexible video verification configurations, dynamically adjusting verification intensity based on user risk profiles.

4. **Identifying Potential Fraud Threats**  
**Dingxiang Dinsight** helps businesses conduct risk assessments, anti-fraud analysis, and real-time monitoring to enhance the efficiency and accuracy of risk control. With an average processing speed of under 100 milliseconds for daily risk control strategies, Dinsight supports multi-party data integration and risk strategy refinement. Combined with the **Xintell Smart Model Platform**, it automatically optimizes risk control strategies based on logs and data mining, uncovering potential threats. Leveraging network association and deep learning technologies, the platform standardizes the complex processes of data processing, feature engineering, and machine learning, providing end-to-end modeling services from data preparation to deployment.

 AI Scams Targeting the Elderly (Part 2): Scammers Impersonating Celebrities and Law Enforcement

Recently, the Rio de Janeiro police successfully dismantled a fraud gang that used AI technology to forge the voice of Brazil’s national team coach, **Dorival Thomaz Junior**, 

Recently, the Rio de Janeiro police successfully dismantled a fraud gang that used AI technology to forge the voice of Brazil’s national team coach, **Dorival Thomaz Junior**, deceiving multiple players and coaches.

#  Fake “Emergency Fundraising” Request

In this case, scammers impersonated Dorival Thomaz Junior and called Brazil's national team players and coaches, asking them to donate money for flood relief in a specific area. During the calls, the fake "Dorival Thomaz Junior" used an urgent and serious tone to stress the importance and immediacy of taking action to support the disaster-stricken region.

![2025011002.jpg](https://cdn.aisecurius.com/images/e3a/e3a5d1d8-2cc3-4c0b-8898-7970b4ae59a6.jpg)

When receiving the calls, the victims did not realize the voice might have been AI-generated, believing they were genuinely speaking with their coach. Particularly in the sports world, where team members and coaches often maintain close relationships, communication is typically carried out via phone calls or private messages. Consequently, this sudden donation request did not arouse suspicion. Many victims complied with the instructions and transferred donations to the designated account provided by the scammers.

Brazilian sports commentator **Bueno** also received a donation request call from the fake “Dorival Thomaz Junior.” However, sensing something unusual, Bueno quickly contacted Dorival Thomaz Junior through alternative means. The real coach clarified that he had never initiated any donation request and warned that it was a scam.
Bueno stated, *“I found it strange, so I reached out to Dorival. He told me it was a scam. Then, he reported it to the police and asked me to warn as many people as possible.”* This timely action helped prevent more potential victims from falling prey to the scam.
After several months of investigation, Brazilian police eventually apprehended the fraud gang. Police revealed that the scammers had been active since 2022, using AI to generate celebrity voices as part of their schemes. Their victims included not only Brazil’s national team players and coaches but also other prominent figures in the sports community.
Authorities noted that this new type of scam is highly covert and poses significant harm to public trust. By leveraging AI to impersonate well-known individuals, scammers turned familiar communication methods into tools of deception, drastically increasing the success rate of their schemes.
This case highlights the application of AI technology in fraud and serves as a reminder to remain calm, rational, and cautious when confronted with urgent financial requests. By forging celebrity voices, scammers exploit familiarity and trust to manipulate victims, emphasizing the need for skepticism and verification in such situations.



# Where Was Dorival Thomaz Junior’s Voice Stolen?


Dorival Thomaz Junior’s voice was leaked primarily due to his frequent public exposure through events, match broadcasts, media interviews, and social media platforms. These channels are typically highly public and widely disseminated, making it easy for scammers to extract sufficient audio material and use AI technology to replicate his voice.  

![2025011001.jpg](https://cdn.aisecurius.com/images/d67/d674653e-593d-4316-8792-7cfd9cda5b85.jpg)

1. **Public Events and Match Broadcasts**  
   During public sports events, coaches often participate in pre-match and post-match interviews, press conferences, and discussions about the team and matches. In these situations, Dorival’s voice is often broadcast or recorded. For instance, in televised national team matches or interviews, journalists and media outlets may record the coach's statements and distribute them widely, forming a publicly accessible audio archive. These audio files are often publicly available online, even as video clips or part of media reports, which makes them highly susceptible to scammers extracting and using them.  

2. **Media Coverage and Interviews**  
   Sports journalists regularly interview the Brazilian national team coach, especially during matches, pre-game preparations, and post-game analyses. Dorival may appear on various media platforms for special interviews or live sessions. These interviews, broadcast worldwide, become public audio or video content. Scammers collect these publicly available audio files and use AI technology to simulate and replicate his voice, allowing them to impersonate him in fraudulent activities.  

3. **Social Media Platforms**  
   The coach may use social media to convey messages to fans or the public, and these platforms can also be sources for scammers to steal his voice. Particularly in live streams or video content, the coach’s statements and tone can be recorded and distributed in real-time, providing material for scammers to mimic.  

4. **Sports Videos and Documentaries**  
   Many media companies and sports platforms create documentaries or feature stories about teams and coaches, often including extended interviews, training footage, or match commentary. In these video materials, the coach’s voice is clearly captured. Due to the high frequency of these videos being played, scammers can easily locate and extract clips of Dorival’s voice, using AI technology to generate synthetic audio nearly indistinguishable from the original.  

In conclusion, the voices of celebrities, particularly in sports, are highly public. Public appearances, media coverage, and social media content are part of their daily lives, making their voices widely recorded and distributed. These audio records are easily accessible, providing scammers with sufficient material to create synthetic voices.  


#  How to Prevent Phone Scams Impersonating Celebrities?


To prevent scams where celebrity voices are stolen and used fraudulently, especially scams using AI-generated voices to impersonate celebrities or loved ones, the Dingxiang Security Cloud Business Intelligence Center advises the public to stay vigilant and adopt multiple verification measures.  

1. **Call Back to Verify Identity**  
   Hang up the call and contact the person through a known, reliable method (e.g., a saved number for a friend or official contact information from a verified source). Do not call back the number provided by the caller, as scammers may use spoofed caller IDs.  

2. **Contact the Individual or Celebrity's Agent Through Other Means**  
   If you receive a call claiming to be from the Brazilian national team coach, try reaching out directly to the official social media accounts of the Brazilian football team or the coach’s agency to verify the claim, avoiding any decisions prompted by the scammer.  

3. **Consult Friends and Family**  
   If you receive a call or message asking for donations or transfers, consult with trusted friends or family members first. Check whether they have received similar requests or if such incidents have occurred. If uncertain, ask them to help verify the authenticity of the call.  

4. **Be Alert to AI Forgery Technology**  
   When receiving a call, even if the voice sounds familiar, be skeptical if there are inconsistencies in tone, fluency, or emotional expression. AI-generated voices may lack subtleties such as natural intonation and genuine emotion.  


Brazil National Team Coach’s Voice Forged by AI; Players and Coaches Scammed

**According to the "Fifth Survey Report on the Living Conditions of Elderly People in Urban and Rural China," as of 2021:**  

**According to the "Fifth Survey Report on the Living Conditions of Elderly People in Urban and Rural China," as of 2021:**  
- 36.6% of elderly individuals can use smartphones.  
- 40.4% cannot use smartphones.  
- 23.0% do not own a smartphone.  
Among those who can use smartphones:  
- 74.7% use mobile chatting apps.  
- 18.1% use mobile payments.  
- 9.6% book medical appointments online.  
- 7.1% use ride-hailing apps.  


# Elderly Frequently Fall Victim to AI Fraud Impersonating Relatives  


**AI Fake Relative Claiming Car Accident**  
In March 2024, Hanyang District Police in Wuhan received reports from a bank where two elderly customers requested early withdrawals from fixed-term deposits, raising suspicions of telecom fraud. The two seniors, unrelated and handling business at the same bank, each received phone calls allegedly from their grandsons.  

![2024121001.jpg](https://cdn.aisecurius.com/images/895/8951a077-3809-4312-97de-74b346e61d69.jpg)

One woman was told her grandson had injured someone in a conflict and was at the police station resolving the issue. The other man was told his grandson was in a car accident and needed immediate compensation for settlement. Both firmly believed the voices sounded exactly like their grandsons. After confirming with family members, the police discovered neither grandson faced such situations. This new type of fraud uses AI technology to clone a relative’s voice, fabricating realistic emergency scenarios.  

**AI Fake Relative Asking for Tuition Fees**  
A septuagenarian, Mr. Zhang, received a call from someone claiming to be his son abroad, saying, “Dad, I’m in trouble overseas and need money urgently.” The voice sounded identical to his son’s, prompting Mr. Zhang to immediately wire his savings. Days later, when his actual son called, it was revealed that the earlier call was a scam using AI voice synthesis.  

![2024121002.jpg](https://cdn.aisecurius.com/images/d2f/d2f5f9fa-8489-43af-a644-da5adc661cb4.jpg)

Another victim, Mrs. Wang, received a video call seemingly from her daughter, studying abroad. The video displayed her daughter’s face, claiming she needed urgent funds for tuition and living expenses. Believing her “daughter’s” plea, Mrs. Wang transferred the money, only to later discover that her real daughter had not initiated the video call. The scammer had used AI face-swapping technology to replicate her daughter’s face, orchestrating a false familial crisis.  


# Why Scammers Succeed  


1. **Information Gathering**  
Fraudsters collect personal data, including names, addresses, contacts, and family details, through social media, public databases, or leaked records.  

2. **AI Voice and Face Simulation**  
Fraudsters use AI to clone voices and generate realistic appearances in video calls, making impersonation highly convincing.  

3. **Fabricating Urgent Scenarios**  
They create scenarios designed to evoke anxiety and urgency, such as claiming accidents or crises that require immediate financial assistance.  

4. **Exploiting Emotional Vulnerability**  
Many elderly individuals, experiencing reduced social circles and feelings of loneliness, have strong emotional dependence on family members. When faced with seemingly authentic appeals from relatives, they lower their guard, especially in emergencies. Emotional stress often impairs judgment, causing them to act hastily without verifying the situation.  

Additionally, many seniors lack awareness of advanced AI technologies like face-swapping or voice synthesis. Unfamiliarity with these technologies leaves them unable to discern fraud in situations where scammers replicate familiar voices or faces seamlessly.  

#  How to Effectively Prevent Such Scams?

The white paper on financial AIGC audio and video anti-fraud（[Click here to download the white paper for free](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)], jointly released by the[ Bank of Communications and Dingxiang](https://www.aisecurius.com/blog/article/131), provides a detailed overview of the characteristics, attack processes, and prevention measures for fraud involving AI-based "face-swapping" and "voice cloning."

![2024121315.jpg](https://cdn.aisecurius.com/images/c77/c775074d-b75b-4f45-91f9-918028633ea2.jpg)

In this new type of telecom fraud, the realistic audio and video generated by AI make it difficult for victims to detect abnormalities through visual or auditory means. Therefore, even after verifying the other party's identity through phone calls or video chats, elderly individuals must remain highly vigilant when faced with urgent financial requests. In transactions involving money, multiple methods must be used to confirm the other party's identity to avoid making hasty decisions that result in falling for a scam. The Dingxiang Defense Cloud Business Security Intelligence Center recommends the following effective measures for the elderly to prevent such fraud:

#### **1. Redial Video or Phone Calls**
- **Action:** When receiving a suspicious call or video, stay calm and do not immediately respond to any requests. You can pretend there is a poor signal, hang up, and then call back using another known and secure contact method to confirm whether the request genuinely came from your relative or friend.
- **Goal:** Avoid directly responding to fraudulent information.

#### **2. Set Up a "Family Verification Mechanism"**
- **Action:** Predefine a "safety word" or "challenge question" known only to you and your close relatives or friends for identity verification. If you receive a suspicious call, ask the other party to provide this keyword.
- **Warning:** Inability to answer correctly or attempts to evade the question should be taken as warning signs of fraud. Hang up immediately and verify the identity through secure means.

#### **3. Consult Trusted Relatives or Friends**
- **Action:** When encountering urgent financial requests, first check with other family members or trusted people to confirm whether they are aware of the situation or have also received similar calls.
- **Benefit:** Seeking advice from trusted individuals can help alleviate pressure and avoid being misled by fraudulent information.

#### **4. Report to the Police**
- **Action:** Call the local police hotline (110) to report the fraudulent situation in detail.
- **Purpose:** Help the police quickly identify and take action against fraudsters to prevent others from becoming victims.

These preventive measures not only help individuals resist AI-related scams but also remind everyone to stay calm and vigilant when dealing with unexpected financial demands to avoid falling into criminal traps.


# How Platforms Can Strengthen Prevention


Fraudsters combine advanced technology with psychological manipulation to target victims precisely. To counter AI forgery scams, efforts from all sectors of society are necessary. These efforts include technical defenses, public education, and emotional support to build a robust defense wall against fraud. Especially, short-video platforms must adopt multiple technologies and measures to detect and block accounts operated by fraudsters at their source and effectively identify forged celebrity videos.

#### **1. Detecting Abnormal Devices**
- **Technology:** Dingxiang’s device fingerprinting technology records and compares device behavior, identifying malicious devices such as virtual machines, proxy servers, and emulators.
- **Functionality:** It analyzes whether devices exhibit abnormal behavior, such as frequent account logins, frequent IP address changes, or device attribute modifications inconsistent with user habits, helping trace and identify fraudsters’ activities.

#### **2. Identifying Abnormal Account Activity**
- **Monitoring:** Focus on scenarios such as accounts logging in from different locations, changing devices, or reactivating dormant accounts.
- **Solution:** Use Dingxiang's atbCAPTCHA technology to verify users during high-risk operations continuously. This ensures persistent identity verification throughout sessions, helping to quickly differentiate between legitimate users and bots and intercept fraudulent behavior in real time.

#### **3. Preventing Fake Face-Swapping Videos**
- **Detection Solution:** Dingxiang's full-chain, panoramic facial security threat detection system verifies devices, facial information, image forgery, user behavior, and interaction status across multiple dimensions.
- **Key Capabilities:** It identifies over 30 types of malicious attacks, including injection attacks, liveness forgery, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, and emulators. The system can automatically block operations involving forged videos, fake facial images, or abnormal interactive behavior. It also allows flexible configuration of video verification intensity, providing a dynamic mechanism to adjust authentication levels based on user risk.

#### **4. Detecting Potential Fraud Threats**
- **Risk Assessment:** Dingxiang's Dinsight platform helps businesses perform risk evaluation, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control.
- **Key Features:** The average processing speed of Dinsight's daily risk control strategies is under 100 milliseconds. It supports multi-source data integration and self-iteration mechanisms using mature indicators, strategies, models, and deep learning techniques. The Xintell intelligent modeling platform, integrated with Dinsight, provides a standardized solution for data processing, feature engineering, model development, and deployment. It enhances the automation of fraud detection strategies by mining risk control logs and identifying potential risks, optimizing security strategies for known risks in one click.


AI Fraud Targets the Elderly (Part 1): Forged Faces and Voices of Relatives Lead to Trust

The solution offers robust protection for ticketing information on B2C websites and official apps by detecting and preventing unauthorized data extraction. Its key features include:

With the arrival of the winter aviation season, Sichuan Airlines has introduced a new flight network layout, including the launch and restoration of several routes. Additionally, the airline has strengthened its operational system's security, focusing on preventing seat hoarding and malicious web-scraping attacks. These measures have significantly enhanced platform protection, ensuring stable operations and a seamless booking experience for customers.

During the winter season, Sichuan Airlines launched and resumed numerous domestic and international routes, including new routes such as Chengdu Tianfu–Changde, Chengdu Tianfu–Shiyan, Harbin–Taiyuan–Beihai, and Xichang–Wuhan–Haikou, among others. They also restored 16 routes, such as Beijing–Mianyang–Xishuangbanna and Chengdu Tianfu–Tengchong. With these changes, the airline now operates over 260 domestic routes and 30 international and regional routes, connecting more than 130 cities.
![2025010301.jpg](https://cdn.aisecurius.com/images/423/42342e9a-bf44-4d16-ba18-b52f3dd667b8.jpg)
As passenger demand and routes grow, challenges like seat hoarding and malicious web scraping have emerged. Such activities disrupt normal flight operations, cause economic losses, and negatively impact user experiences. To address these issues, Sichuan Airlines implemented Dingxiang's aviation anti-crawling solution.

The solution offers robust protection for ticketing information on B2C websites and official apps by detecting and preventing unauthorized data extraction. Its key features include:

1. **atbCAPTCHA**: Real-time detection and interception of malicious accounts and bot activities during user registration, login, and searches.
2. **Device Fingerprinting**: Continuous monitoring to identify and mitigate risks like code injection, emulators, and jailbroken devices, ensuring system stability.
3. **Dinsight Real-Time Decision Engine**: Behavior-based analysis to identify potential threats, securing all transactions.

These technologies enable Sichuan Airlines to effectively combat seat hoarding and prevent economic losses caused by web scraping. They enhance order processing efficiency and ensure a smooth, secure booking experience for customers. Continuous monitoring and data analysis allow the airline to adapt its security measures to evolving threats.

To date, Dingxiang's anti-crawling solutions have supported major airlines like Air China, Travelsky, and Kunming Airlines, promoting safer and healthier development within the aviation industry.

Sichuan Airlines Deploys Dingxiang Aviation Anti-Crawling Solution

To further safeguard the App's security, Soul has chosen to use Dingxiang App Hardening to enhance platform security, protect user privacy, and ensure compliance with regulatory requirements.

Soul is a social platform based on interest maps, designed with gamified features for Gen Z users. It aims to empower users to express their self-worth and hopes to break through geographical and other social limitations, becoming a virtual playground that helps users discover, establish, and maintain social relationships.

As China’s first social network platform where all users interact via virtual avatars, Soul has become a highly active and sticky social community formed through a massive volume of user-generated content (UGC), decentralized content distribution mechanisms, and a series of self-developed technologies. It leverages virtual avatar-based social interaction, open mobile social networking based on interest maps, and a range of other self-developed technologies to create a platform for highly active user engagement.

To continuously protect the user social experience, accurately identify and punish undesirable social behaviors, Soul has innovatively developed features such as “AI-assisted identification,” “network begging behavior detection,” and “fake persona identification.” The platform has developed and launched more than 430 anti-fraud strategies, significantly improving its fraud detection capabilities. According to reported data, the number of cases originating from Soul App has consistently declined, with cases in the third quarter decreasing by 13% compared to the previous quarter, demonstrating the effectiveness of the platform's governance efforts.

![2024122601.PNG](https://cdn.aisecurius.com/images/551/5518015a-f9a2-4a3d-b9fd-4e800b5e38de.PNG)

In response to the widespread issue of “fake personas” on social platforms, such as impersonating a ‘police officer’ or a ‘divorced single parent,’ Soul quickly identifies and locks down potentially risky user groups. Through a multi-dimensional examination approach, including but not limited to: abnormal age and location, suspicious personas, high-risk professions, potential victim traits, and image theft, the risk control model continuously monitors these indicators to ensure the platform’s environment remains safe and authentic.

Soul has also launched the “Instant Fake Persona Risk Alert” feature. Since its launch, it has achieved an accuracy rate of over 95% and a coverage rate of 60%. In the past month alone, over 570,000 alerts have been sent. Many fake personas, when constructing their identity narratives, often show inconsistencies or contradictions with real situations. By accurately capturing these identity contradictions and classifying the relevant users, this feature significantly improves image recognition efficiency within the platform, enhances the user experience, and creates a community environment more aligned with users’ expectations.

To further safeguard the App's security, Soul has chosen to use Dingxiang App Hardening to enhance platform security, protect user privacy, and ensure compliance with regulatory requirements. First, App hardening enhances the app's ability to resist attacks, preventing hackers from intruding, tampering, or reverse engineering, thereby ensuring the security of user data. Second, Soul uses hardening technology to protect its anti-fraud features, preventing fraudsters from bypassing protective mechanisms and ensuring the effectiveness of fake identity and online scam prevention. Furthermore, App hardening strengthens data encryption and access control, ensuring that user privacy remains protected and improving platform security and user trust. Additionally, Soul App needs to comply with China’s Cybersecurity Law and Data Security Law, and App hardening helps meet these compliance requirements.

Dingxiang App Hardening is an adaptive hardening solution based on graph neural network technology. By deeply analyzing and extracting features from the app's code, it automatically selects the optimal encryption, obfuscation, and compression methods, increasing the difficulty of reverse analysis, and effectively preventing the app from being cracked, copied, or repackaged.

- **Deep Analysis and Feature Extraction**: The application code is comprehensively scanned and analyzed to extract key features and automatically select the best hardening approach.
- **Multi-Layer Encryption and Obfuscation**: Combining multi-layer encryption and obfuscation techniques to increase code complexity, preventing reverse engineering and cracking.
- **Compression Technology**: Optimizing the code size while maintaining security, improving the app’s performance.
- **Real-Time Protection**: A dynamic monitoring and real-time defense mechanism detects and blocks potential security threats promptly.

App hardening supports the high-activity, high-engagement needs of social platforms, protecting UGC content and decentralized social mechanisms, further ensuring the long-term healthy development of the platform.

Soul Chooses Dingxiang App Hardening

Recently, a video posted by an account named “Ban Hua XXX” showed a well-known doctor, Zhang Wenhong, promoting a protein bar. However, 

Recently, a video posted by an account named “Ban Hua XXX” showed a well-known doctor, Zhang Wenhong, promoting a protein bar. However, media investigations revealed that while the lip movements and voice in the video resembled Zhang Wenhong’s, the promotion was not done by him. The video was confirmed to be a deepfake, created using AI technology. Screenshots from users indicated that 1,266 units of the protein bar had been sold. Zhang Wenhong later told the media that there were multiple such fraudulent accounts, which frequently changed, making it challenging to report them.
![2024121901.png](https://cdn.aisecurius.com/images/cb9/cb941aa4-8db4-488b-8bdb-60c071d5c5f2.png)

Zhang Wenhong said, “I’ve considered reporting this to the police, but it’s hard to describe an invisible and intangible perpetrator who changes accounts daily and uses only virtual tools. Who can I report? I think this is a violation of consumer rights and should be handled by the relevant authorities.”
![2024121904.png](https://cdn.aisecurius.com/images/f5c/f5c06f0a-9ba8-4e61-8448-204d83addc34.png)
On December 15, the WeChat platform released a statement: “Recently, media reports have highlighted cases of improper marketing using AI-generated impersonations of celebrities. We have taken swift and strict action against accounts that use AI to impersonate well-known individuals for malicious marketing or to gain traffic improperly, and we have launched a special crackdown campaign. So far, 532 pieces of content have been removed, and 209 accounts have been closed. Moving forward, we will increase our efforts to combat violations related to ‘using AI to impersonate public figures for marketing purposes.’”


# How Did Scammers Create Fake Videos of Dr. Zhang Wenhong?


AI-based deepfake technology enables the creation of highly convincing fake images or videos. AI voice cloning can also mimic a target’s voice with high accuracy. These AIGC (AI-Generated Content) fraud techniques have become significant risks in daily life, threatening public safety. Using “deepfake” and “voice cloning” technology, scammers can impersonate celebrities, family members, company executives, or government officials, conducting telecom scams that result in substantial financial losses for victims.

![2024121315.jpg](https://cdn.aisecurius.com/images/c77/c775074d-b75b-4f45-91f9-918028633ea2.jpg)

The **[Financial AIGC Audio and Video Anti-Fraud White Paper](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)**, [published by the Bank of Communications and Dingxiang](https://www.aisecurius.com/blog/article/131), outlines the attack process of [deepfake and voice cloning fraud.](https://www.aisecurius.com/blog/article/132)

### AIGC “Deepfake” Attacks  
1. Fraudsters obtain high-resolution ID photos of victims through purchased identity information.  
2. These photos are modified using tools like Photoshop to include background elements.  
3. AI tools are then used to animate the photos, creating fake videos of blinking or head movements.  
4. These fake videos are injected into apps to bypass identity verification systems.

### AIGC “Voice Cloning” Attacks  
1. Fraudsters gather voice samples from victims through scam call recordings.  
2. AI tools are used to generate synthetic audio mimicking the target’s voice.  
3. These audio files are used in real-time or injected attacks to exploit identity verification systems.


# Fake Celebrity Advertisement Fraud Process  


Fraudsters use AI to create fake celebrity advertisements, with a process that involves data collection, audio and video synthesis, advertisement distribution, and persuading victims to transfer money. This method leverages public trust in celebrities and advanced technology to enhance scam credibility.

#### Step 1: Data Collection  
Fraudsters collect large amounts of publicly available information about a target celebrity, such as photos, videos, interviews, and social media posts. These materials are used as input for AI systems to create fake content.

#### Step 2: Fake Audio and Video Creation  
With sufficient data, scammers use AI to synthesize audio and video. This technology allows them to replicate the celebrity’s voice and facial expressions accurately, producing fake advertisement videos. These videos are often indistinguishable from authentic ones.

#### Step 3: Fake Advertisement Distribution  
The fake videos are distributed across various online platforms, including short video apps, social media, and group chats. These platforms often target elderly users, who are more vulnerable. Fraudsters use precise marketing techniques to reach specific demographics. Due to the realistic appearance and sound of the videos, victims are easily convinced of their authenticity.

#### Step 4: Building Trust and Authority  
To increase success rates, scammers create a sense of urgency or use language that inspires trust. This reinforces the legitimacy of the advertisement.

#### Step 5: Inducing Purchases  
When victims believe the content is genuine, they are directed to purchase links or customer service hotlines in the video. The products either do not exist or are substandard, resulting in financial loss for the victims.

# How Ordinary People Can Identify Fake Celebrity Advertisements  

Fraudsters use AI technology to create fake celebrity advertisements through a systematic process, from data collection to audio-video synthesis, advertisement distribution, and inducing victims to transfer money. These scams exploit public trust in celebrities and advanced techniques to enhance credibility. So, how can ordinary people spot fake celebrity advertisements?

1. Be skeptical of shocking or too-good-to-be-true claims, especially involving celebrities. Don’t forget common sense and logic.  

2. Carefully check for unnatural features in the content, such as strange blinking patterns, mismatched audio, or distorted body parts, which may indicate deepfakes.  

3. Verify the claims by cross-checking with reliable news media and websites. Don’t trust content without fact-checking it first.  

4. Consult friends, family, or colleagues for advice before making decisions or purchases. Conduct in-depth research and don’t take the promoter’s word as gospel, even if it’s someone like Elon Musk.  

5. Avoid commenting, sharing, or even clicking on unverified social media posts. Engaging with accounts impersonating celebrities increases the risk of being scammed.  

6. Never share personal details with unverified individuals, websites, or apps.  


#  How Platforms Can Prevent Fake Celebrity Advertisements  


To address fake celebrity advertisements and information on social media platforms, the **Dingxiang Business Security Intelligence Center** recommends accurately identifying fake accounts as the first step in effectively preventing fraud.

1. **Identifying Abnormal Devices**  
   Dingxiang’s device fingerprinting technology records and compares device fingerprints to distinguish between legitimate users and potential fraudsters. It uniquely identifies devices and detects maliciously controlled devices, such as virtual machines, proxy servers, or emulators. The technology analyzes device behaviors, such as multi-account logins, frequent IP address changes, and alterations in device attributes that deviate from user habits, helping track and identify fraud activities.

2. **Identifying Abnormal Account Operations**  
   Abnormal activities, such as logging in from different locations, device changes, phone number changes, or sudden activity in dormant accounts, require frequent verification. Continuous identity verification during sessions is essential to maintain consistency. Dingxiang’s **atbCAPTCHA** accurately distinguishes between humans and bots, identifies fraud in real-time, and intercepts abnormal behavior.  

3. **Preventing Deepfake Videos**  
   Dingxiang’s full-chain facial security threat detection solution uses multi-dimensional verification, including device environment, facial information, image authenticity, user behavior, and interaction status. It quickly identifies over 30 types of malicious attacks, such as injection attacks, fake videos, face spoofing, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, and emulators. Upon detecting fake videos, fake face images, or abnormal interactions, the system can automatically block the operation. The solution also allows dynamic configuration of video verification intensity, balancing user experience for normal users with stricter verification for suspicious ones.

4. **Detecting Potential Fraud Threats**  
   Dingxiang’s **Dinsight Engine** assists enterprises with risk assessment, fraud detection analysis, and real-time monitoring to improve efficiency and accuracy in risk control. Dinsight processes routine risk control strategies in under 100 milliseconds and supports multi-source data configuration and retention. It leverages mature metrics, strategies, and deep learning to enable performance monitoring and self-iteration for risk control. Combined with **Xintell**, its intelligent modeling platform, it optimizes security strategies for known risks and identifies potential risks through log analysis and data mining. Xintell standardizes the process from data processing and feature engineering to model deployment, providing end-to-end modeling services.  

Using AI to Impersonate “Zhang Wenhong” to Sell Health Products, Over 1,200 Victims Scammed

Mr. Li received a text message claiming to be from his workplace leader, asking him to add a social media account for further communication. 

Mr. Li received a text message claiming to be from his workplace leader, asking him to add a social media account for further communication. After becoming "friends," the fraudster requested an urgent transfer of funds, claiming it was needed for financial turnover, with a promise of repayment later. Initially doubtful, Mr. Li was convinced when the scammer initiated a video call. The person in the video appeared to be his “leader,” which made Mr. Li lower his guard. He then made three separate transfers totaling 950,000 yuan. 

![2024121001.jpg](https://cdn.aisecurius.com/images/895/8951a077-3809-4312-97de-74b346e61d69.jpg)

Later that day, during a conversation with friends, Mr. Li found the situation suspicious and decided to report it to the police. After receiving his report, the police quickly coordinated with the bank to freeze the involved funds through emergency measures, successfully recovering 830,000 yuan and preventing greater losses. Thanks to the swift action of the authorities, most of Mr. Li’s money was retrieved.

In Another Case, an Elderly Woman Is Scammed by AI Voice Cloning。Ms. Li received a call from someone claiming to be her "younger brother." The caller stated that he had changed his phone number and asked her to add him on WeChat, promising to visit her soon. A few days later, the "brother" called again, claiming he was detained after a fight and needed money for compensation. He asked Ms. Li for financial assistance. Believing the story, she prepared 70,000 yuan in cash. The scammer then requested an additional 50,000 yuan. Though Ms. Li began to feel suspicious, she still decided to report the incident to the police.

After investigation, the police confirmed this was a case of fraud using AI voice cloning technology. The scammers had used realistic voice synthesis to mimic Ms. Li’s brother’s voice, tone, and speaking habits, completely earning her trust. The police initiated collaborative operations and successfully tracked down the suspect, Mr. Wang, in another province. Mr. Wang was apprehended and has been detained on fraud charges while the case undergoes further investigation.


#  Why the Victims Fell for the Scams


In both cases, the victims failed to recognize the signs of manipulated audio and video. Mr. Li trusted the video call, unaware that the video could have been digitally altered. Similarly, Ms. Li was convinced by the realistic voice mimicry, which replicated not only her brother’s voice but also his speech patterns, eliminating her doubts. This trust led her to provide significant financial support without hesitation.

# Fraudsters Use AI to Identify Target Victims

The **"[Financial AIGC Audio and Video Anti-Fraud White Paper](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)"** [published by the Bank of Communications in collaboration with Dingxiang highlights the processes and characteristics of AI-based scams, ](https://www.aisecurius.com/blog/article/131)such as "face-swapping" and "voice cloning."
![2024121315.jpg](https://cdn.aisecurius.com/images/c77/c775074d-b75b-4f45-91f9-918028633ea2.jpg)
AI face-swapping technology is now widely used in scams. Fraudsters use this technology to superimpose a target’s face onto their own, creating synthetic videos or photos that make scams appear highly credible. This technology can "pass as real," trapping victims without their knowledge.

Additionally, AI is used to screen potential victims. Scammers no longer rely on blind attempts but analyze personal information shared on online platforms. Using AI, they precisely select targets and design personalized scam strategies. This approach allows them to execute scams more efficiently and covertly, significantly increasing the success rate of their fraudulent activities. 


# How to Effectively Prevent AI-Based Scams?


The **["Financial AIGC Audio and Video Anti-Fraud White Paper](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)"** calls on industries to establish a comprehensive defense system covering the entire cycle, a[ll scenarios, and the full fraud chain to combat AI-enabled fraud. For individuals, the following measures can effectively help prevent scams](https://www.aisecurius.com/blog/article/132):
![2024121309.jpg](https://cdn.aisecurius.com/images/402/4024349f-5457-4891-bfda-2475edc84b62.jpg)

#### **1. Redial Videos or Calls**
When receiving suspicious calls or video requests, stay calm and avoid responding immediately. Use the excuse of "poor signal" to hang up and then call back using a verified, safe contact method. Confirm whether the request truly came from your friend or family member, and avoid responding directly to potentially fraudulent messages.

#### **2. Set Up Family Verification Mechanisms**
Predefine a "safety word" or a "challenge question" that only you and your family members know. If you receive a suspicious call, ask the caller to provide this keyword. Failure to respond correctly or attempts to evade the question should raise red flags. In such cases, hang up immediately and verify the person's identity through secure channels.

#### **3. Consult Friends or Family**
In cases involving urgent financial requests, consult other friends or family members to verify the situation. Ask if they are aware of any similar calls or messages. This prevents you from making decisions based on stress or misinformation. Reaching out to trusted contacts can help you assess the situation more accurately.

#### **4. Report to Police**
Dial your local emergency number, such as 110 in China, to report the fraud to authorities in detail.

These measures can help individuals resist AI-based scam threats. They also serve as a reminder to stay calm and vigilant when encountering sudden financial requests, avoiding falling into the traps set by fraudsters.


# Platforms Must Strengthen Anti-Fraud Measures


To address AI-based scams, society must work together by combining technical solutions, public education, and emotional support. This joint effort will help prevent more people from becoming victims. Short-video platforms, in particular, need to adopt multiple technologies and measures to detect fraudsters and identify fake audio and video content from the source.

#### **1. Identifying Abnormal Devices**
**Dingxiang Device Fingerprinting** enables the identification of fraudulent activities by tracking and comparing device fingerprints. This technology assigns unique IDs to devices and detects malicious activities such as virtual machines, proxy servers, simulators, and abnormal behaviors like multi-account logins, frequent IP address changes, and device attribute modifications. These measures help track and identify fraudsters effectively.

#### **2. Detecting Abnormal Account Operations**
Abnormal activities such as logging in from different locations, device changes, new phone numbers, or sudden activity from dormant accounts require increased verification. Continuous identity verification during sessions is crucial. **Dingxiang atbCAPTCHA** accurately distinguishes between humans and bots, detecting fraudulent behavior in real time and blocking abnormal activities.
![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)
#### **3. Preventing Deepfake Videos**
**Dingxiang Full-Link Facial Security Threat Detection** uses multi-dimensional information, including device environment, facial data, image verification, user behavior, and interaction status, to conduct intelligent authentication. It quickly identifies over 30 types of malicious activities, such as injection attacks, live spoofing, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, and emulators. Upon detecting fake videos, false facial images, or abnormal interactions, it automatically blocks operations. The system also provides flexible configurations for video verification strength, enabling user-friendly verification for legitimate users while applying enhanced validation for suspicious ones.

#### **4. Uncovering Potential Fraud Threats**
**Dingxiang Dinsight** helps enterprises conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving risk control efficiency and accuracy. With an average processing speed of 100 milliseconds for daily strategies, Dinsight supports multi-source data integration and storage, leveraging metrics, strategies, and deep learning-based models. Combined with the **Xintell Intelligent Modeling Platform**, it automatically optimizes security strategies, mines potential risks based on logs and data, and customizes fraud prevention strategies for various scenarios. By standardizing processes from data handling to model deployment, this end-to-end platform enhances the effectiveness of risk control.

Fraudster Uses AI to Impersonate a Company Leader and Scams Mr. Li Out of 830,000 Yuan

The *Financial AIGC Audio-Video Anti-Fraud White Paper* is divided into seven chapters, primarily covering the risks of audio-video fraud posed by AIGC,

Recently, at the "Digital Intelligence Empowerment: Financial Technology Driving High-Quality Development in Finance" forum hosted by Bank of Communications, the ***Financial AIGC Audio-Video Anti-Fraud White Paper*（[Click here to download](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)）** was released. Jointly authored by Bank of Communications, Dingxiang Technology, and RealAI, this white paper systematically explores the risks and challenges brought by AIGC (AI-generated content), focusing on the audio-video fraud challenges faced by the financial industry. It serves as a reference for financial institutions to enhance their capabilities in identifying and preventing AIGC-related fraud. The white paper provides detailed insights into professional AIGC audio-video detection technologies and products designed to combat AIGC audio-video fraud.


#  AIGC Audio Forgery Detection Technology  


The core objective of AIGC audio detection technology is to improve accuracy and reliability in the face of continuously advancing audio forgery techniques, focusing on audio quality, voiceprint features, and spectrum analysis.  

![2024121702.png](https://cdn.aisecurius.com/images/afe/afe917ed-f23d-44e6-b24e-412e5217c198.png)
**Audio Quality:** Forged audio often exhibits quality anomalies, such as noise, distortion, or other imperfections that affect clarity. Since the process of generating synthetic audio typically introduces these unnatural elements, they serve as important clues for detecting fake audio.  

**Voiceprint Features:** Each person’s unique vocal characteristics, shaped by their physiological traits like vocal cord structure and speaking habits, create distinct voiceprints. AIGC-generated audio often lacks the personalized nuances of human speech, tending to be overly regular and mechanical in aspects like tone and pitch, which forms the basis for voiceprint detection.  

**Spectrum Analysis:** Spectrum analysis converts audio signals from the time domain to the frequency domain to examine frequency components. In audio forgery detection, AIGC-generated audio often exhibits unnatural features in high or low-frequency ranges, such as irregular frequency distribution. These anomalies can be revealed through spectrogram analysis.  

AIGC audio detection technology combines multi-level feature fusion, adversarial training, and temporal modeling to maintain high-precision detection performance even in the face of various generation techniques and complex noise interference. Looking ahead, the integration of multimodal information and more advanced deep learning technologies is expected to further improve the detection of forged audio.


# AIGC Image Forgery Detection Technology  


The primary task of AIGC image forgery detection is to determine whether an image was generated or tampered with by artificial intelligence. Evidence of forgery is typically reflected in visual artifacts, digital signal anomalies, model fingerprints, facial priors, and violations of physical imaging principles. 

![2024121703.png](https://cdn.aisecurius.com/images/613/6134fdfb-effa-4dcc-8d53-d2116ae853cb.png) 

**Visual Artifacts:** AIGC image generation may produce unnatural visual effects caused by algorithmic limitations, insufficient training data, or computational constraints. For example, generated images might display unclear or distorted details.  

**Digital Signal Anomalies:** Forged images may show irregularities in frequency, noise, or color statistics. In the frequency domain, AI-generated images may exhibit specific patterns, such as artifacts introduced by upsampling operations. In the noise domain, generated images often have noise patterns distinct from those in real images. These signal-level anomalies can be analyzed to detect forged images.  

**Model Fingerprints:** The process of AI-generated images leaves unique “model fingerprints,” determined by the architecture, training data, and parameter settings of the generative model. Specialized detection models can identify these fingerprints to determine whether an image was generated by a specific model.  

**Facial Priors:** In generated facial images, anomalies such as misaligned eye positions, irregular pupil shapes, or unnatural nose structures may appear. These issues often indicate the image was AI-generated.  

**Physical Imaging Principles:** Real images adhere to physical rules of light propagation, reflection, and refraction. In contrast, AIGC-generated images may exhibit phenomena that violate these rules, such as unnatural shadow directions or incorrect perspective relationships.  

Methods for detecting forgery clues in AIGC images can be broadly divided into handcrafted approaches and representation learning. Handcrafted methods design specific features based on an understanding of forgery mechanisms to identify traces of manipulation. Representation learning, on the other hand, leverages deep learning to handle the complex and evolving challenges of image forgery in an automated and efficient manner. Although representation learning offers adaptability and efficiency, handcrafted methods remain irreplaceable in certain scenarios due to their transparency and interpretability.  


# AIGC Video Forgery Detection Technology  


The core objective of AIGC video detection technology is to accurately determine whether video content is AI-generated or tampered with. Compared to image forgery detection, video forgery detection involves not only clues present in static images but also dynamic features such as temporal visual artifacts, audiovisual inconsistencies, and the naturalness of motion trajectories.  
![2024121704.png](https://cdn.aisecurius.com/images/b9c/b9c9a9c4-7c6b-4314-9c77-ba7466268a9f.png)
Clues to forgery in AIGC videos include static image forgery clues as well as temporal and dynamic variations in the video. **Temporal visual artifacts** refer to unnatural transitions between consecutive frames in a video, often manifested as frame jumps or discontinuous changes. **Audiovisual inconsistencies** arise when the visual and auditory information in a video are not perfectly synchronized, particularly in generated videos where sound fails to align with the visuals. **The naturalness of motion trajectories** is another key clue. In AI-generated videos, object movements may appear rigid, jerky, or physically implausible, such as abrupt speed changes or abnormal acceleration.  

Modeling methods for detecting forgery clues in AIGC videos can be categorized into handcrafted approaches and representation learning.  

**Handcrafted Feature Modeling:**  
This approach relies on traditional "feature engineering," focusing on video-specific attributes. For example, techniques such as optical flow and trajectory tracking can analyze object motion trajectories within a video to detect unnatural movement or abrupt speed changes. Additionally, analyzing spatiotemporal consistency between video frames, such as the coherence of color, brightness, and texture across consecutive frames, can help identify potential forged areas. Synchronization between audio and video content is another critical factor to verify whether a video has been edited.  

**Representation Learning:**  
This method leverages deep learning models to automatically learn effective feature representations from large-scale video datasets. Compared to representation learning in image forgery detection, video forgery detection must pay particular attention to temporal dimensions and motion continuity. By combining 3D convolutional neural networks (CNNs) and recurrent neural networks (RNNs), it becomes possible to capture both spatial and temporal features in videos, enabling effective detection of forgery clues. Additionally, multimodal learning and cross-modal consistency verification techniques are employed to ensure synchronization between visual and auditory content in videos, further enhancing the precision of forgery detection.  

 Feature Association Analysis Based on Knowledge Graphs  

Knowledge graph-based AIGC feature association analysis, as a graph-structured data model, utilizes deep learning, feature extraction, and fingerprint recognition technologies to structurally describe relationships between entities (such as people, events, and locations) and uncover potential complex connections between different data objects. This approach not only effectively identifies potential fraud groups involved in AIGC-generated content but also reveals behavioral patterns and links among multiple fraudulent individuals.  
![2024121705.png](https://cdn.aisecurius.com/images/e31/e31412a5-356e-4d37-8e5c-65dc1e8127c3.png)
By constructing knowledge graphs based on AIGC features, **community detection algorithms** can analyze high-density groups of nodes within the graph, identifying individuals belonging to the same fraud group. These nodes often share similar characteristics or behaviors, such as using the same AIGC generation tools or displaying similar forgery features. **Association inference** helps identify suspicious behavior patterns, uncovering consistent behavioral traits across multiple nodes within the same community, such as frequent use of identical audio forgery or face-swapping techniques.  

Furthermore, **group expansion and path tracing** can reveal the propagation paths of fraud networks by tracking relationships between nodes. For instance, by identifying nodes associated with multiple fake accounts, potential group members can be further uncovered, thereby revealing the complete fraud network. This knowledge graph-based analysis method provides robust technical support for precise targeting of AIGC-related fraud.  

The*****Financial AIGC Audio-Video Anti-Fraud White Paper***（[Click here to download](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)）**is divided into seven chapters, primarily covering the risks of audio-video fraud posed by AIGC, typical attack methods of AIGC audio-video fraud, its impact on financial business, anti-fraud solutions for AIGC audio-video, technical implementations for combating AIGC audio-video fraud, typical application scenarios, and outlook and recommendations.  

White Paper: Four Key Technologies for Effectively Detecting Face-Swapping and Voice Cloning  

The "AIGC Audio-Video Anti-Fraud White Paper" calls on industry enterprises to establish a defense system covering the entire lifecycle, all scenarios,

At the recent "Digital Intelligence Driving, Open Win-Win: Financial Technology Empowering High-Quality Financial Development" forum hosted by the Bank of Communications, the **"AIGC Audio-Video Anti-Fraud White Paper**"（[Click here to download](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)） was released. Li Zhaoning, General Manager of the Bank of Communications' Network Finance Department and Director of the Financial Technology Innovation Research Institute, provided a detailed interpretation of the white paper.

![2024121305.png](https://cdn.aisecurius.com/images/2b0/2b08039f-f01d-4ff2-943c-559f499197e1.png)

The white paper, co-authored by the Bank of Communications, Dingxiang Technology, and RuiLai Intelligence, aims to systematically explore the risk challenges posed by AIGC technology applications, focusing on the audio-video fraud challenges faced by the financial industry, and providing reference for financial institutions to enhance their ability to detect and prevent AIGC fraud. The white paper elaborates on the attack risks, details, and security defense solutions of AIGC "face-swapping" and "voice-cloning," as well as professional multimodal AIGC audio-video anti-fraud technology.


# The Rise of Audio-Video Fraud with AIGC


The rapid development of AIGC technology has brought unprecedented security risks to identity verification systems in the financial industry. Especially in widely used remote facial recognition and voiceprint recognition systems, black and gray markets have leveraged AIGC tools to carry out "face-swapping" and "voice-cloning" attacks, which have become a serious threat.

![2024121315.jpg](https://cdn.aisecurius.com/images/c77/c775074d-b75b-4f45-91f9-918028633ea2.jpg)

**AIGC "Face-Swapping" Attack:**  
Remote facial recognition systems are widely used in various financial services, such as account opening, credit card applications, and claims processing, as an identity verification method to ensure user security. The system workflow includes multiple steps such as face data collection, liveness detection, quality testing, and face comparison. However, black and gray market actors can easily create fake videos using AIGC tools. By customizing client ROMs or hijacking cameras, they inject counterfeit victim videos into the client during the face collection process. Through forged actions such as "blinking" and "shaking head," they bypass liveness detection, thus overcoming identity verification measures and launching an attack.

**AIGC "Voice-Cloning" Attack:**  
Voiceprint recognition systems, as another identity verification method, also face challenges from AIGC. Voiceprint recognition works by collecting users' voice information and analyzing their voiceprint features to verify identity, commonly used in telephone banking and mobile financial services. Black and gray market actors can acquire victim voice samples, such as recordings from phone scams, and use AIGC tools to generate counterfeit audio. During the voice collection process, attackers can play back the fake audio to bypass the voiceprint comparison system, successfully launching an identity verification attack.

Both of these AIGC attack methods share a common characteristic: attackers use AIGC technology to generate highly realistic fake audio-video content, enabling them to carry out fraud without traditional identity verification methods. As AIGC technology continues to evolve, financial institutions urgently need to strengthen anti-fraud technologies, enhance system security, and adopt more stringent detection and prevention measures to ensure accurate and secure customer identity verification, thus addressing increasingly complex cybersecurity challenges.


# Financial Institutions Need to Establish a Comprehensive Defense System


The  **"AIGC Audio-Video Anti-Fraud White Paper**"（[Click here to download](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)）calls on industry enterprises to establish a defense system covering the entire lifecycle, all scenarios, and the full chain to respond to the complex threats posed by AIGC and safeguard the security of financial services.

1. **Multimodal AIGC Audio-Video Anti-Fraud Technology**  
The technical defense system is one of the core elements in preventing AIGC fraud. Financial institutions should promptly adopt advanced multimodal audio-video anti-fraud technologies, integrating techniques such as audio-video forgery detection, device correlation analysis, and user behavior recognition to improve the accuracy and timeliness of fraud detection.  

![2024121308.jpg](https://cdn.aisecurius.com/images/17e/17e84eb5-787f-48ff-ad28-dd282b69adfb.jpg)
The white paper provides a detailed introduction to AIGC audio-video anti-fraud technology:  
- **AIGC Video Detection:** This technology analyzes inconsistencies in image segmentation and motion between frames in a video to detect forged video content. It can identify subtle anomalies in the video, such as distorted facial expressions or disjointed body movements, helping financial institutions effectively recognize counterfeit videos.  
- **AIGC Voice Detection:** Voice forgery is a common method in AIGC fraud. Fraudsters use deep forgery technology to imitate others' voices. Voice detection technology can effectively distinguish real speech from generated speech by identifying abnormal audio spectra and complex noise interference, thus preventing identity impersonation and fraud during phone or video calls.  
- **AIGC Feature Correlation Analysis Based on Knowledge Graphs:** This technology identifies and tracks potential fraudulent groups in AIGC-generated content. By deeply analyzing the connections and coordinated behavior of multiple fraudsters, it reveals their modus operandi. Combining graph computing and knowledge graphs, it can provide financial institutions with more comprehensive anti-fraud strategies to identify and combat cross-platform, cross-scenario fraud activities in real time.

This multimodal anti-fraud technology demonstrates its powerful anti-fraud capabilities in multiple financial service scenarios. In credit business, the system analyzes customer identity verification data, credit history, and social behaviors to prevent the risks of false information and forged identities, improving loan approval efficiency and accuracy while reducing the occurrence of non-performing loans. For credit card services, combining user consumption habits, transaction history, device fingerprints, and location data helps effectively prevent credit card fraud and identity theft. In the payment settlement domain, by analyzing transaction behaviors, device characteristics, IP addresses, and payment frequencies in real time, it can quickly identify and prevent fraudulent activities, reducing financial losses.

**2. Strengthening Training for Financial Practitioners**

Strengthening training for financial practitioners is crucial in addressing AIGC deep forgery fraud. It not only improves employees' ability to detect fraud but also helps financial institutions improve their risk management systems, reducing potential losses.

Firstly, training can significantly enhance practitioners' risk response capabilities. By systematically learning, employees can identify forged audio-video content, understand the basic principles of deep forgery technology, and recognize subtle signs of forgery in facial recognition and voice patterns, thus effectively preventing identity theft and financial losses. Additionally, training helps employees become proficient in using anti-forgery technologies, such as face-swap detection tools and audio detection tools, to promptly flag suspicious activities.

Secondly, training strengthens risk and decision management. Through regular training, practitioners better understand the risks of AIGC fraud and can implement strict verification measures in customer identity checks and transaction monitoring. Trained employees can identify potential risk signals and take timely actions, such as freezing accounts or conducting secondary verification, effectively reducing fraud losses.

Finally, training also enhances customer trust and institutional reputation. Trained employees can provide professional anti-fraud advice, address customer concerns, and improve customer satisfaction. Furthermore, training ensures employees comply with laws and regulations, guaranteeing that anti-fraud measures are compliant, thereby reducing regulatory risks and maintaining the financial institution’s public image.

**3. Optimizing Legal and Compliance Management**

Establishing a scientific management system is key for financial institutions to address AIGC audio-video fraud.

Firstly, financial institutions need to build a unified risk management framework, comprehensively assessing AIGC risks from prevention, monitoring, to emergency response, ensuring that all departments and personnel can efficiently address fraud threats. By integrating technology, process controls, and personnel management, financial institutions can cover the full lifecycle of risk management, enhancing overall anti-fraud capabilities.

Secondly, compliance is a critical component of responding to AIGC fraud. Financial institutions must ensure that their anti-fraud measures comply with data protection and privacy laws, taking legal action to hold perpetrators accountable and protect legal rights. As AIGC technology develops, financial institutions need to continuously update compliance management to ensure the effectiveness of anti-fraud measures.

Additionally, by integrating data across departments and systems, financial institutions can monitor customer transaction behavior in real-time and quickly detect potential fraud. The system can identify unusual activities, such as atypical login locations or large transactions, and trigger corresponding anti-fraud measures, improving detection and response capabilities. Real-time monitoring technology and intelligent decision engines can help institutions dynamically assess the authenticity of audio-video content and quickly identify forged content.

Finally, regular anti-fraud drills and simulation tests can improve emergency response capabilities, helping financial institutions identify vulnerabilities in their defense systems, optimize technologies and processes, and ensure efficient handling of actual fraud cases to reduce risks.

**4. Improving the Legal System and Regulatory Mechanism**

At the legal and regulatory level, it is necessary to improve relevant laws and regulations and strengthen enforcement against AIGC fraud to ensure that emerging risks from AIGC can be effectively addressed amidst technological innovation and development. Strengthening regulatory oversight is critical. Government regulatory bodies should establish stronger oversight mechanisms, tracking and reviewing the technological deployments of financial institutions to ensure the safety of financial services and prevent the misuse of AIGC technology.  

Establishing inter-departmental collaboration mechanisms is also essential. A cross-departmental collaboration mechanism should be set up to strengthen rapid response and joint efforts in combating AIGC fraud cases in the financial industry.

The  **"AIGC Audio-Video Anti-Fraud White Paper**"（[Click here to download](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)）is divided into seven chapters, covering the risks of audio-video fraud brought by AIGC, typical AIGC audio-video fraud attack methods, the impact of AIGC audio-video fraud on financial services, AIGC audio-video anti-fraud solutions, the implementation of AIGC audio-video anti-fraud technology, typical business scenarios, and outlook and recommendations.

Bank of Communications White Paper Reveals the Attack Process of “Face-Swapping and Voice Cloning” Fraud

Recently, the **"Digital Intelligence Driven, Open Win-Win: Fintech Empowering High-Quality Financial Development"** forum was held in Shanghai, hosted by the Bank of Communications. 

Recently, the **"Digital Intelligence Driven, Open Win-Win: Fintech Empowering High-Quality Financial Development"** forum was held in Shanghai, hosted by the Bank of Communications. Over 200 guests from government agencies, financial institutions, and technology enterprises attended the event, focusing on the deep integration of artificial intelligence and data elements. Discussions centered on AI large model applications in the financial sector, exploring fintech development trends and practical opportunities, providing clear directions and strategic recommendations for the industry's future. **Qian Bin**, a Party Committee member and Vice President of the Bank of Communications, delivered a keynote speech at the forum.

![2024121302.png](https://cdn.aisecurius.com/images/e30/e306876c-80e4-4197-80aa-750ee85efe18.png)

At the event, the **"White Paper on Financial AIGC Audio-Video Anti-Fraud"**（[Click here to download the white paper for free](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)] was unveiled. Jointly authored by the Bank of Communications, Dingxiang Technology, and RealAI, the white paper systematically discusses the risk challenges brought by AIGC technology applications, focusing on audio-video fraud issues in the financial industry. It provides references to help financial institutions enhance their capabilities in identifying and preventing AIGC-related fraud. **Li Zhaoning**, General Manager of the Bank of Communications’ Network Finance Department and Director of the Fintech Innovation Research Institute, gave an in-depth explanation of the white paper during the forum.

# The Prevalence of AIGC Leads to Rising Audio-Video Fraud

The  **"White Paper on Financial AIGC Audio-Video Anti-Fraud"**（[Click here to download the white paper for free](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)] elaborates on the development process of AIGC technology. Since the release of **MidJourney** and **ChatGPT** in 2022, AIGC has made rapid advancements in just two years, evolving from text-to-text and text-to-image to deep forgery and text-to-video generation. While AIGC's widespread adoption offers significant innovative potential, it has also quietly facilitated various fraudulent activities.

![2024121305.png](https://cdn.aisecurius.com/images/2b0/2b08039f-f01d-4ff2-943c-559f499197e1.png)

#### Examples of Global AIGC Fraud:
- **January**: A Hong Kong company was defrauded of HKD 200 million during a virtual meeting.
- **April**: An American password management company employee received a fake call impersonating the CEO.
- **May**: CCTV exposed an AI-generated video of an old classmate, leading to a woman being scammed for RMB 400,000.
- **July**: Hangzhou police disclosed that fraudsters used AI to forge facial recognition and log into others’ accounts.
- **September**: Fake audio of Lu Moumou from Hefei’s "Three Sheep Company" circulated widely online.
- **October**: A cloned voice of Lei Jun went viral during the National Day holidays.
- **Recent Months**: Fraudsters have increasingly exploited AIGC to impersonate relatives, celebrities, and judicial officials, particularly targeting the elderly in scams.

AIGC-based **facial swapping** can generate highly deceptive forged images or videos, while **voice cloning** enables the creation of voice content that closely mimics the target. Audio-video fraud using AIGC has become a significant risk for financial institutions, challenging business security, customer trust, and system stability.

#### Industry Reports Highlight Growing Risks:
- A **BioCatch** report revealed that nearly **70%** of respondents believe criminals are becoming more proficient at using AI for financial fraud.
- The **World Economic Forum (WEF)** estimated that the number of forged videos is growing at an annual rate of approximately **900%**.
- A **Deloitte** report indicated that **deepfake incidents** in the financial industry surged by **700%** in 2023, and projected that the U.S. financial sector could lose up to **$40 billion by 2027**, with a compound annual growth rate of **32%**.

# Typical Fraud Methods Involving AIGC Audio-Video Technology

With **facial swapping** and **voice cloning**, malicious actors can engage in fraudulent activities such as mass account creation, account login, fund transfers, and loan applications. These activities directly result in significant financial losses for customers, pose severe threats to financial operations, and disrupt the normal functioning of national financial services.

![2024121306.png](https://cdn.aisecurius.com/images/685/68540ab7-a221-4791-aee1-cff24122b837.png)

#### 1. AIGC Facial Swapping Attacks:
- **Step 1**: Obtain high-definition facial images of the victim through purchased identity information.
- **Step 2**: Use Photoshop to prepare facial images with appropriate backgrounds.
- **Step 3**: Employ AIGC tools to animate the victim's photos, creating fake videos with blinking or head movements.
- **Step 4**: Inject the forged videos into apps to target identity authentication systems.

#### 2. AIGC Voice Cloning Attacks:
- **Step 1**: Collect voice data from victims, often through telephone scam recordings.
- **Step 2**: Use AIGC tools to generate fake audio mimicking the victims’ voices.
- **Step 3**: Exploit the forged audio through playback or injection attacks to compromise identity authentication systems.

Challenges and Requirements for Financial Anti-Fraud Measures

AIGC audio-video fraud introduces new risks for financial operations, providing fresh attack methods for malicious actors. These developments present new challenges for security defenses and raise the bar for anti-fraud measures in the financial sector.


# Financial Institutions Must Establish Comprehensive Defense Systems


The  **"White Paper on Financial AIGC Audio-Video Anti-Fraud"**（[Click here to download the white paper for free](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)] urges industry players to establish a defense system that encompasses the **entire lifecycle, all scenarios, and all chains** to address the complex threats posed by AIGC and ensure the security of financial operations.

![2024121309.jpg](https://cdn.aisecurius.com/images/402/4024349f-5457-4891-bfda-2475edc84b62.jpg)


#### Technical Level
Introduce multimodal AIGC audio-video fraud detection and authentication technologies to accurately identify forged content and respond to AIGC fraud in real-time.

#### Personnel Level
Enhance the training of financial professionals to improve their ability to recognize and respond to AIGC-related fraud, helping institutions better manage risks.

#### Management Level
Establish a robust management system, optimize legal and compliance frameworks, and improve the anti-fraud capabilities of financial institutions.

#### Legal and Regulatory Level
Refine laws and regulations, strengthen enforcement, and enhance supervisory efficiency to ensure that AIGC technology fosters financial innovation while effectively preventing potential risks and misuse.



# Multimodal AIGC Audio-Video Anti-Fraud Technology


A **technical defense system** is one of the core elements for preventing AIGC fraud. Financial institutions should adopt **advanced multimodal audio-video anti-fraud technologies**, leveraging methods such as forged content detection, device correlation analysis, and user behavior identification to enhance detection accuracy and real-time responses.

![2024121308.jpg](https://cdn.aisecurius.com/images/17e/17e84eb5-787f-48ff-ad28-dd282b69adfb.jpg)
#### Detailed Overview of Anti-Fraud Technologies in the White Paper:

1. **AIGC Video Detection**  
   This technology identifies forged video content by analyzing inconsistencies in frame segmentation and motion. It detects subtle anomalies in videos, such as **distortions in facial expressions** or **incoherent body movements**, helping financial institutions identify fake videos effectively.

2. **AIGC Voice Detection**  
   Voice forgery is a common AIGC fraud method where criminals use deepfake technology to mimic others’ voices. Voice detection technology distinguishes real from generated voices by identifying abnormalities in voice spectrums and complex noise patterns, preventing impersonation and fraud through calls or video conferences.

3. **Knowledge Graph-Based AIGC Feature Correlation Analysis**  
   This technology tracks potential fraud rings by analyzing connections and collaborative behaviors among multiple fraud entities. By combining **graph computation** and **knowledge graphs**, financial institutions can gain comprehensive anti-fraud strategies, enabling timely identification of cross-platform and cross-scenario fraud activities.

![2024121311.jpg](https://cdn.aisecurius.com/images/8c4/8c44a596-f243-46b2-b7ac-e57d4127d56d.jpg)
#### Real-World Applications
Many financial institutions have successfully managed AIGC-related fraud risks by combining technical defenses and manual interventions.

- **Example Case**: A bank's remote video service faced fraudulent attacks from malicious actors using AIGC technology. In response, the bank adopted a combination of technological defenses and manual intervention.  
   - Its **AIGC audio detection system** improved the detection rate of forged audio to **99%-100%**, significantly reducing fraud risks in remote video services.  

By integrating these systems, financial institutions can create a more secure environment, safeguarding their operations and customer trust.


# Three Major Initiatives to Prevent AIGC Fraud  


In the future, AIGC audio-video fraud is expected to become more intelligent, multidimensional, automated, and personalized. Continuous attention, early planning, and proactive responses are required. The **"White Paper on Financial AIGC Audio-Video Anti-Fraud"** proposes three major initiatives, calling for collaboration among governments, industries, enterprises, and institutions to establish a comprehensive, multi-layered defense system to address new challenges posed by AIGC, ensuring the security and reliability of the financial system.

![2024121312.png](https://cdn.aisecurius.com/images/092/092884bc-cc85-40ae-919e-7c905a2a09e9.png)

#### Strengthen the Compliance System  
- Continuously update laws and regulations to effectively suppress new risks and respond to emerging challenges.  
- Refine standards for the transparency and interpretability of AIGC applications, guiding enterprises to prevent audio-video forgery risks in technological applications, business processes, and management mechanisms.

#### Innovate and Advance Technology  
- Develop more advanced fraud detection models to combat increasingly sophisticated attacks.  
- Build automated countermeasures to block large-scale fake content generated by fraud rings.  
- Provide customized identity verification and risk warning services to ensure robust risk prevention while maintaining an optimal customer experience.

#### Build a Healthy Ecosystem  
- Foster collaboration among academia, industry, and practical applications to improve the efficiency of technology transfer and rapidly defend against evolving challenges.  
- Promote industry-wide sharing to support enterprises and institutions in adjusting operational strategies, algorithms, and risk assessments.  
- Strengthen public awareness campaigns to enhance anti-fraud literacy, laying a foundation for trust and order within the industry.

![2024121315.jpg](https://cdn.aisecurius.com/images/c77/c775074d-b75b-4f45-91f9-918028633ea2.jpg)

The **"White Paper on Financial AIGC Audio-Video Anti-Fraud"**（[Click here to download the white paper for free](https://cdn.dingxiang-inc.com/files/10354/AIGC%20Anti-fraud.pdf)）is divided into seven chapters, covering the following topics:  
1. Risks of audio-video fraud brought by AIGC.  
2. Typical attack methods in AIGC audio-video fraud.  
3. Impacts of AIGC audio-video fraud on financial operations.  
4. Solutions for combating AIGC audio-video fraud.  
5. Implementation of AIGC audio-video anti-fraud technologies.  
6. Typical business scenarios.  
7. Future outlook and recommendations.

Bank of Communications, Dingxiang Technology, and RealAI Release the "White Paper on Financial AIGC Audio-Video Anti-Fraud"

Fraud cases targeting the elderly are increasing. In the latest incidents, scammers used AI technology to mimic the voices and faces of relatives, successfully deceiving elderly victims into 

Fraud cases targeting the elderly are increasing. In the latest incidents, scammers used AI technology to mimic the voices and faces of relatives, successfully deceiving elderly victims into transferring money intended for living expenses and tuition.
![2024121001.jpg](https://cdn.aisecurius.com/images/895/8951a077-3809-4312-97de-74b346e61d69.jpg)

Mr. Zhang, a septuagenarian, has been living alone since his wife passed away years ago. His son works abroad and rarely visits. One day, Mr. Zhang received a call from an unfamiliar number. The voice on the other end was unmistakably that of his son. "Dad, I’m in trouble overseas and need money to resolve the issue," the voice said, filled with urgency and tension. Concerned for his son, Mr. Zhang quickly decided to send his hard-earned savings to “help.” A few days later, during a call with his real son, Mr. Zhang discovered he had been scammed. The entire incident was orchestrated by fraudsters using AI voice-cloning technology to replicate his son’s voice.

Another heartbreaking incident involved Mrs. Wang, whose daughter is studying abroad. The two often communicate via video calls. One day, Mrs. Wang received a video call request that displayed her daughter’s familiar face. At first, Mrs. Wang suspected nothing, as she was accustomed to such calls. During the conversation, however, the “daughter” claimed to be in urgent need of tuition and living expenses. Driven by maternal concern, Mrs. Wang quickly transferred the money. Later, she discovered that her real daughter had not initiated the call. The “daughter” on the screen was actually a scammer using AI face-swapping technology to superimpose her daughter’s face onto their own.


# Unveiling the Fraudsters’ Techniques


1. **Information Gathering**  
   Scammers begin by collecting basic information about the elderly victim through social media, public records, or leaked databases. This includes names, addresses, contact details, and information about their relatives.
![2024121002.jpg](https://cdn.aisecurius.com/images/d2f/d2f5f9fa-8489-43af-a644-da5adc661cb4.jpg)
2. **AI Voice and Face Cloning**  
   Next, they use AI technology to clone voices or generate realistic facial replicas. In video call scams, fraudsters employ AI face-swapping tools to create highly convincing fake visuals, making the "relative" in the video appear flawless and believable.

3. **Creating Urgent Scenarios**  
   Scammers craft scenarios designed to provoke anxiety, such as claiming to have been in an accident or facing an emergency abroad. These scenarios compel the elderly, who are emotionally attached to their relatives, to act quickly without questioning authenticity.

4. **Emotional Manipulation**  
   By continuously increasing the sense of urgency and exploiting the elderly victim's emotional dependency on their relatives, fraudsters push victims to make hasty decisions—often transferring money without sufficient verification.


# Why Are Elderly People Prone to Fraud?


1. **Emotional Dependence on Family**  
   Many elderly individuals experience declining physical abilities and shrinking social circles, intensifying feelings of loneliness. This emotional void often creates a strong desire and dependence on familial bonds. When scammers exploit these familial ties, the elderly are more likely to lower their guard and trust the "relative" on the call or video.
![2024121003.jpg](https://cdn.aisecurius.com/images/5e1/5e1c54c5-8498-4f41-b594-7eb4475ac296.jpg)

2. **Lack of Awareness About AI Technology**  
   Most elderly people are unfamiliar with advanced AI technologies like voice cloning and face-swapping, which have reached highly convincing levels of realism. Due to limited tech literacy, they struggle to distinguish between genuine and fraudulent interactions, especially when scammers create seamless impersonations of their loved ones.

3. **Emotion-Driven Decision Making**  
   In urgent scenarios fabricated by scammers, the elderly often let their emotions override rational judgment. Under pressure, they may rush to decisions without thoroughly verifying the information, leading to financial losses.


# How to Effectively Prevent Such Scams


In these sophisticated telecom frauds leveraging AI-generated audio and video, it is challenging to detect anomalies through direct observation. Even after verifying identity via calls or video, the elderly should maintain vigilance in response to urgent financial requests. Multiple identity verification methods should be employed to avoid falling victim to such scams. Dingxiang Defense Cloud’s Security Intelligence Center recommends the following preventive measures:

- **Redialing Calls or Videos**  
   Upon receiving a suspicious call or video, stay calm and avoid responding immediately to any requests. Use an excuse like poor signal to hang up, then reconnect via known, secure contact methods to confirm the requestor’s identity, bypassing the scammer’s manipulation.

- **Setting Up Family Verification Mechanisms**  
   Prearrange a “safe word” or “challenge question” known only to you and your family. When suspicious calls arise, ask for the keyword. Failure to answer correctly or evasive behavior can be a warning sign. Hang up immediately and verify the identity through secure channels.

- **Consulting Trusted Relatives**  
   For emergency monetary requests, consult with other relatives or trusted contacts to see if they are aware of the situation. This can help alleviate pressure and provide a clearer perspective, reducing the risk of deception.

- **Reporting to Authorities**  
   Contact local authorities by dialing 110 to report the fraud in detail.

These measures not only help individuals combat AI-driven fraud but also remind everyone to stay calm and vigilant during unexpected financial requests, reducing the likelihood of falling into scammers’ traps.


# Platforms Should Strengthen Prevention


Fraudsters are combining technology and psychological manipulation to control victims with precision. To combat AI-based fraud, collective societal efforts are essential, using technical, educational, and emotional support to create a robust barrier against scams. Particularly, short-video platforms must adopt multiple techniques to identify fraudulent accounts and detect manipulated audio and video content at the source.

1. **Detecting Anomalous Devices**  
   **Dingxiang Device Fingerprinting** can distinguish legitimate users from potential fraudsters by analyzing unique device identifiers. It detects malicious devices such as virtual machines, proxies, and emulators. By monitoring behaviors like multi-account logins, frequent IP changes, and abnormal device activity, it helps trace and identify fraud.

2. **Identifying Abnormal Account Behavior**  
   Accounts showing unusual patterns like remote logins, device swaps, or sudden activity after dormancy should undergo frequent verification. Continuous session-based identity checks are critical. **Dingxiang atbCAPTCHA** quickly differentiates between humans and bots, identifying fraudulent activity in real time and blocking it effectively.
![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)

3. **Preventing Fake Videos Using Face-Swapping**  
   **Dingxiang's Full-Link Panorama Face Security Solution** employs intelligent verification through device environment analysis, facial recognition, image authenticity checks, user behavior monitoring, and interaction states. It identifies over 30 types of malicious activities, such as live spoofing, image forgeries, camera hijacking, and system modifications. Suspicious videos or interactions can be blocked instantly, with flexible verification configurations balancing user experience and fraud detection.

4. **Uncovering Potential Fraud Risks**  
   **Dingxiang Dinsight** supports risk assessment, fraud analysis, and real-time monitoring for enterprises, enhancing risk control efficiency and accuracy. With processing speeds under 100 milliseconds, it integrates multi-source data and applies advanced strategies and models for fraud detection. Paired with the **Xintell Intelligent Model Platform**, it optimizes security strategies for known risks and uncovers hidden threats, offering standardized data processing, feature extraction, and one-click fraud control for various scenarios.

Two Elderly People Received "Relative Calls" That Turned Out to Be AI Fraud

In late March 2024, an elderly woman, Mrs. Zhang, from a residential area in Xiaogan City, received a phone call that left her in a panic. On the other end of the line was the voice of her “grandson

In late March 2024, an elderly woman, Mrs. Zhang, from a residential area in Xiaogan City, received a phone call that left her in a panic. On the other end of the line was the voice of her “grandson,” Xiao Wan, urgently saying, “Grandma, please help me, I’ve caused trouble!” Mrs. Zhang, distressed, immediately inquired about the situation. Xiao Wan explained that he had broken a friend's phone, and after a conflict with the friend, he was required to pay 20,000 yuan in compensation, or the friend would report him to the police and have him jailed. Xiao Wan begged his grandmother to help, saying that he could not collect the money himself, and asked her to give the cash to a friend who would come to collect it.

![2024090401.jpeg](https://cdn.aisecurius.com/images/b3d/b3d8baf3-0f9f-4c53-9667-17d4ad081f59.jpeg)

Driven by anxiety and concern, Mrs. Zhang withdrew her 20,000 yuan in savings and handed it over to a man wearing a mask, as instructed. Afterward, Mrs. Zhang, still worried, called Xiao Wan to check if the issue had been resolved, only to be shocked when he expressed surprise and said that he had never been in any trouble and had not asked his grandmother for money. At that moment, Mrs. Zhang realized she had been scammed and immediately reported the incident to the police.

The police quickly retrieved surveillance footage and identified three suspects involved in the case. The investigation revealed that the core member of the fraud ring, Zhou, had used illegal communication software to contact the “boss” and directed Yuan A and Yuan B to carry out the scam. Zhou used AI technology to clone Xiao Wan’s voice and made phone calls to Mrs. Zhang to carry out the fraud. Meanwhile, Yuan B, pretending to be Xiao Wan’s friend, took the cash from Mrs. Zhang and quickly left the scene.

The prosecuting attorney explained that the criminals used AI technology to generate a voice almost identical to Xiao Wan's and used this voice remotely to communicate with Mrs. Zhang, making her trust that the person on the phone was truly her grandson. As a result, Mrs. Zhang was scammed out of her 20,000 yuan savings.

After a thorough investigation, the police confirmed the method of the crime and identified all the involved parties. Following court proceedings, the three suspects were found guilty of fraud and sentenced. Zhou was sentenced to one year in prison and fined 10,000 yuan, while Yuan A and Yuan B were sentenced to 10 months in prison and fined 5,000 yuan each.


# AI Technology Becomes a New Tool for Crime


The criminals used AI voice-swapping technology to successfully mimic Mrs. Zhang's "grandson’s" voice, creating a phone call that was nearly impossible to distinguish as fake. As a result, even if one attempts to verify identity via phone or video calls, heightened vigilance is necessary when confronted with urgent requests for funds. When money transactions are involved, it is crucial to use multiple methods to confirm the other party's identity to avoid making a hasty decision and falling victim to fraud.

The Dingxiang Defense Cloud Business Security Intelligence Center advises the elderly to take the following preventive measures:

1. **Video or Phone Redial**: When receiving suspicious calls or video requests, stay calm and do not immediately respond to any demands. Hang up the phone citing poor signal and then call back using other known and secure contact methods to confirm whether the request came from a real family member.

2. **Set up a Family Verification Mechanism**: Pre-arrange a "safety word" or "challenge question" with your family that only you and your loved ones know. If you receive a suspicious call, ask the caller for the safety word. If the caller cannot provide the correct word or avoids the question, it is a clear sign of a scam. Hang up immediately and verify the identity through secure means.

3. **Consult Trusted Family and Friends**: In case of urgent requests for money, check with other family members or trusted individuals to verify the situation or see if they have received similar calls. Consulting someone you trust can help you assess the authenticity of the request and avoid being pressured into making a rash decision.

4. **Report to the Police**: Call the local police at 110 to report the details of the scam.

These precautionary measures not only help individuals defend against AI scams but also remind everyone to remain calm and alert when faced with sudden requests for money, preventing falling into the traps set by criminals.

Fraudster Using AI to Scam Elderly Woman of 20,000 Yuan Sentenced to 1 Year

Fraudsters used sophisticated tactics to impersonate Jack Ma and attempted to lure an elderly woman into purchasing health products. The criminal activity was uncovered due to the timely intervention

Recently, a scam involving AI face-swapping technology was successfully intercepted in a community in Fengtai District, Beijing. Fraudsters used sophisticated tactics to impersonate Jack Ma and attempted to lure an elderly woman into purchasing health products. The criminal activity was uncovered due to the timely intervention of community workers and police officers, preventing any financial loss to the elderly woman.

The incident began when an elderly woman sought help at the community service center and revealed that she had received a phone call from someone claiming to be a "disciple of Jack Ma." The scammer initially gained the woman's trust by sending small amounts of phone credits, and then claimed to be running a poverty alleviation project that promised substantial investment returns. To further reassure the woman, the scammer suggested a video call with "Jack Ma."

![2024112801.png](https://cdn.aisecurius.com/images/954/9544ee9f-7b09-4434-ac9d-f87a46907e11.png)

Using AI face-swapping technology, the fraudster disguised themselves as Jack Ma, appearing on the elderly woman’s phone screen during the video call. Due to the high level of realism of the technology, the elderly woman did not doubt the authenticity of the person in the video, even believing she was speaking directly with Jack Ma. The scammer successfully gained her trust and then asked her to download QQ software for further investment operations.

Fortunately, community workers and police officers quickly identified the scam through detailed questioning, and revealed the truth to the elderly woman during another video call. Upon seeing the police officer on screen, the scammer became visibly panicked and quickly ended the call.

# 
The Role of AI Technology and the Elderly's Psychological Vulnerabilities

1. In this scam, the AI face-swapping technology played a crucial role. The face-swapping technology allowed the scammer to precisely superimpose a real person’s facial features onto another’s, making the “Jack Ma” in the video look nearly identical to the real individual.

![2024112802.png](https://cdn.aisecurius.com/images/d74/d748ee1a-d3b1-4c92-bc75-79ac378706a9.png)

2. Many elderly individuals lack sufficient awareness of internet security, particularly emerging technologies such as AI face-swapping and deepfake technology. As a result, they are more susceptible to scams and are likely to believe they are truly having a video conversation with Jack Ma.

3. Additionally, Jack Ma is widely known and influential in society. By impersonating Jack Ma’s "disciple" and offering a fake “video call” with him, the scammer significantly increased their credibility. The elderly woman, who had never met Jack Ma in person, was easily convinced and lowered her guard.

4. Furthermore, the scammer used a strategy of offering small favors—such as recharging the woman's phone—to build a relationship of trust. Through gradual manipulation, the fraudster gained psychological control over the elderly woman, further lowering her defenses and convincing her of the “investment opportunity.” The elderly woman was eventually drawn into a larger trap set by the scammer.

# Elderly Individuals and Businesses Should Strengthen Fraud Prevention

Police are reminding residents, especially the elderly, to remain vigilant when it comes to preventing fraud. When receiving calls or video calls from strangers, individuals should never blindly trust the identity claimed by the caller, avoid downloading unknown software, and refrain from making any transfers or purchases. Additionally, they should seek help from family members, friends, or community workers to verify the authenticity of the information.

Fraudsters are using a combination of technology and psychological manipulation to precisely control their victims. In the face of AI-driven scams, society needs to build stronger defenses through technology, educational outreach, and emotional support to prevent more people from falling victim. Particularly on short video and social media platforms, multiple technologies and methods need to be employed to identify fraudulent accounts and detect fake celebrity videos from the source.

1. **Identifying Abnormal Devices.** Dingxiang Device Fingerprinting records and compares device fingerprints to distinguish between legitimate users and potential fraud activities. The technology uniquely identifies each device and detects maliciously controlled devices such as virtual machines, proxy servers, and emulators. It analyzes abnormal behaviors such as multiple account logins, frequent IP address changes, and changes in device attributes that do not match user habits, helping track and identify fraudulent activities.

2. **Identifying Accounts with Abnormal Operations.** Login from different locations, device changes, new phone numbers, and suddenly active dormant accounts require frequent verification. Additionally, continuous identity verification during sessions is essential to ensure that the user’s identity remains consistent throughout the interaction. Dingxiang atbCAPTCHA can quickly and accurately distinguish between human and machine operators, precisely identify fraud, and monitor and intercept abnormal activities in real-time.

3. **Preventing Fake Videos with Face-Swapping.** Dingxiang’s full-chain, panoramic facial security threat perception solution uses multidimensional information such as device environment, facial information, image forgery detection, user behavior, and interaction status for intelligent verification. It quickly identifies over 30 types of malicious attacks, including injection attacks, live-body forgery, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, and emulators. After detecting fake videos, fake facial images, or abnormal interaction behaviors, it automatically blocks the operation. It also allows flexible configuration of video verification strength and user-friendliness, implementing a dynamic mechanism to strengthen verification for abnormal users while maintaining normal atbCAPTCHA for legitimate users.

4. **Identifying Potential Fraudulent Threats.** Dingxiang Dinsight helps businesses with risk assessment, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight’s daily risk control strategy processes data in less than 100 milliseconds and supports the customizable integration and storage of multiple data sources. It uses deep learning technologies, established indicators, strategies, and models to enable self-performance monitoring and iterative mechanisms. Paired with the Xintell intelligent model platform, Dinsight can optimize security strategies for known risks, mine potential risks from risk control logs and data, and configure risk control strategies for different scenarios with one click. The platform standardizes complex data processing, mining, and machine learning processes, offering one-stop model-building services from data processing and feature derivation to model deployment.

Additionally, elderly individuals and the public need to understand that AI face-swapping technology can create fake dynamic videos. They should remain alert to high-quality video images and not easily trust the authenticity of video content. In similar situations, do not blindly follow instructions out of fear. Instead, seek help from family members or the police promptly.

Fraudsters Use "AI Face-Swapping" to Impersonate Jack Ma and Lure Elderly Woman into Buying Health Products

Moreover, CAPTCHA, as an important element in user interaction with enterprises, can reflect the style and characteristics of the company. Through customized CAPTCHA appearances, IGUS can convey its

IGUS GmbH is a global company headquartered in Cologne, Germany, specializing in the production and sale of high-performance plastic products. The company is known for developing durable and maintenance-free mechanical components, particularly in the fields of plastic bearings and energy chain systems.

IGUS primarily provides plastic components for a variety of industries, including automation, mechanical manufacturing, automotive, aerospace, and robotics. Its main products and services include energy chain systems used to manage and protect cables and hydraulic hoses in automated and industrial machinery; high-performance plastic bearings for industrial and consumer machinery that require no lubrication and are maintenance-free; sliding bearings suitable for various loads and temperature conditions with low friction and wear-resistant properties; precision linear guides for motion control in automation and manufacturing equipment; and highly flexible cables with high durability, commonly used in high-stress environments in energy chain systems.

IGUS (Shanghai) Energy Chain Systems Co., Ltd. is IGUS GmbH’s wholly-owned subsidiary in China, responsible for pre-sales consulting and after-sales services in the region. To effectively address various risks and further enhance brand awareness, IGUS’s website has chosen Dingxiang atbCAPTCHA for user registration and login verification.
![2024101601.png](https://cdn.aisecurius.com/images/d2d/d2d10806-162e-43fc-8f81-9906e506ab39.png)
Dingxiang atbCAPTCHA, based on AIGC technology, effectively prevents automated programs from performing malicious registrations, bulk login attempts, and password retrieval operations during registration and login processes. It ensures that these actions are being performed by real human users, not malicious programs or hackers. Additionally, by requiring users to enter a CAPTCHA code, it increases the difficulty of account cracking or theft. This enhances user awareness of account security and builds greater trust in the IGUS brand.

Moreover, CAPTCHA, as an important element in user interaction with enterprises, can reflect the style and characteristics of the company. Through customized CAPTCHA appearances, IGUS can convey its brand features and values to users, further shaping a positive corporate image.

Dingxiang atbCAPTCHA offers three personalization management options: custom configuration, logo configuration, and font configuration. Enterprises can freely configure interference items, background images, text, angles, verification methods, and more according to their needs, with an instant preview option.

IGUS GmbH China Website Chooses Dingxiang atbCAPTCHA

The elderly woman, still confused, asked, “He wore a police uniform and could video call me—is this fake?” She explained that the caller accused her of criminal involvement and demanded 

Recently, the "Lei Feng Volunteer Police" in Hunan Provincial Museum spotted an elderly woman behaving nervously outside a mobile phone store while on a FaceTime video call. Sensing a possible scam, they reported it to local police. Officers arrived quickly, finding the woman engaged in a video conversation with someone impersonating a police officer. The caller, dressed in a police uniform, was sternly instructing her to perform certain actions on her phone. The police seized the woman’s phone, identified themselves, and confronted the scammer, who abruptly ended the call.  
![2024112201.png](https://cdn.aisecurius.com/images/9ea/9ea3a029-90d9-42b9-b2b2-e8812e5c425e.png)
The elderly woman, still confused, asked, “He wore a police uniform and could video call me—is this fake?” She explained that the caller accused her of criminal involvement and demanded her cooperation in a "financial investigation." She followed instructions out of fear, believing the scammer’s police identity.  


# What Kind of Scam Was This?  

The scam involved advanced AI face-swapping technology to create a fake video of the scammer posing as a police officer. Combined with FaceTime's video and screen-sharing features, the fraudster manipulated the victim into sharing sensitive data.  

**Key Methods Used by Scammers**:  
- **AI Face-Swapping**: The scammer superimposed a police officer's face onto their own in real-time, with synchronized facial expressions and speech, making the video call highly convincing.  
- **Screen Sharing**: FaceTime’s built-in screen-sharing function enabled the scammer to view the victim's phone activity, including inputting banking details or confidential information.  

### How Did the Scam Work?  

The fraudster accused the victim of criminal activities and demanded cooperation to "verify" her innocence. They instructed her to buy a new phone and SIM card to “assist the investigation.” Through screen sharing, the scammer observed the victim’s phone operations in real-time, capturing her bank account details and personal data.  

### Why Was It Effective?  

1. **Dynamic Interaction**: Unlike static images, AI-driven live video interactions seemed authentic, 
2. 
3. with natural facial expressions and gestures.  
2. **Trust in Law Enforcement**: Seeing a police uniform and receiving strict instructions prompted the victim to associate the caller with genuine authorities.  
3. **Lack of Technical Knowledge**: The victim was unfamiliar with AI face-swapping and FaceTime’s screen-sharing risks, making it difficult to detect the scam.  
4. **Fear and Pressure**: Allegations of criminal involvement induced panic, impairing the victim’s rational judgment and making her compliant.  



# # How Can the Elderly Prevent Such Scams?



1. **Be Wary of “Official Personnel” on Video Calls**  
   Authorities such as police, banks, or government agencies typically do not contact individuals via FaceTime or similar video call apps. If uncertain, verify through official contact numbers instead of trusting caller IDs. Law enforcement does not conduct operations via video calls, nor do they request bank transfers or account details. Always hang up and verify independently if you receive such calls.  

2. **Disable FaceTime’s Sharing Features**  
   If not frequently used, it is advisable to disable FaceTime to prevent misuse. On iPhones, go to “Settings,” locate the “FaceTime” option, and turn it off to eliminate the risk of FaceTime-based scams.  

3. **Refuse Screen Sharing with Strangers**  
   Avoid sharing your screen with strangers on any video app. Be especially cautious if someone claiming to be an official requests access to view your banking operations.  

4. **Watch Out for Phishing Websites and Fake Apps**  
   Do not click on suspicious links. Be extra cautious of websites or pages you are directed to during a screen-sharing session. Always download apps from trusted sources like the Apple Store.  

5. **Protect Personal Information and Privacy**  
   Avoid disclosing sensitive information such as facial data, bank card numbers, OTPs, or passwords to strangers. Stay vigilant to prevent fraudsters from exploiting your personal details.  

6. **Use Technical Safeguards**  
   Tools like **Dingxiang Device Fingerprinting** can detect screen-sharing activities in real time and issue alerts. If screen sharing is activated, it can notify the user about potential risks and advise caution.  

Additionally, the public should remain informed about AI face-swapping technology’s ability to create fake dynamic videos. Exercise skepticism towards seemingly high-quality video images and refrain from entering sensitive data during unfamiliar calls or screen-sharing sessions. In such situations, avoid succumbing to fear and blindly following instructions. Instead, seek assistance from family members or the police.

Elderly Almost Scammed Out of Life Savings via AI Face-Swapping Video Call

On the morning of November 12, the Jin Yuan Police Station of the Economic Development Division of the Yichun City Public Security Bureau, Jiangxi Province, received a distress call from a local bank

On the morning of November 12, the Jin Yuan Police Station of the Economic Development Division of the Yichun City Public Security Bureau, Jiangxi Province, received a distress call from a local bank. Bank staff reported an elderly woman insisting on obtaining a bank loan despite incomplete documentation. She adamantly refused to leave, stating, “I won’t leave without the loan.”

Upon receiving the call, officers promptly arrived at the scene and communicated with the elderly woman. Through patient questioning, they discovered the reason behind her loan application: she wanted to secure 2 million yuan to support her “boyfriend,” Jin Dong, in financing his film project. She claimed to have met this "Jin Dong" on a short video platform, where he had professed his love and shared edited photos of them together. He also explained that he needed funds to support his film production.

#### Police Discovery and Intervention

The officers, suspicious of her story, examined her phone and found that the "Jin Dong" she referred to was an imposter. His identity and the accompanying photos and videos were AI-generated fakes. Despite this, the woman was convinced of his authenticity, believing him to be her idol.  
![2024112203.png](https://cdn.aisecurius.com/images/543/543ea65d-f69f-425e-b9d0-9b1014e61821.png)
The police remained patient, educating her about similar scams reported in the media and demonstrating the principles of AI-generated content. With the assistance of her family, who was contacted to intervene, the elderly woman finally realized the scam and abandoned her loan application.


#  Scammer's Fraud Process


The real Jin Dong is a well-known actor with a positive public image, often portrayed in dramas as a caring and trustworthy character. These roles generate trust, especially among elderly fans, who may blur the line between fiction and reality. Scammers exploited this trust using the following steps:

1. **Creation of AI-Generated Content**  
   Scammers used AI to create realistic short videos mimicking Jin Dong’s appearance, voice, and expressions, making them highly convincing.  

2. **Targeted Distribution via Short Video Platforms**  
   By leveraging platform recommendation algorithms, scammers ensured the fake content reached specific users, quickly capturing their attention.

3. **Building Emotional Connections**  
   Through frequent interactions, such as declarations of affection and filming "short videos together," the fake "Jin Dong" fostered an emotional bond. Presenting the narrative of "a celebrity noticing an ordinary person" further reinforced trust.

4. **Manipulation and Financial Exploitation**  
   Continuous interaction built unwavering trust in the fake identity. Once the victim was emotionally invested, the scammers introduced the financial scheme, claiming it was for a film project, and applied psychological pressure to obtain money.

Elderly individuals are often unfamiliar with AI technology and online scams, making it difficult for them to discern the authenticity of videos or photos. Additionally, loneliness and emotional needs can make them more susceptible to forming false emotional connections, causing them to overlook logical judgment.  


# Platforms Must Strengthen Protections  


Fraudsters are increasingly combining technology with psychological manipulation to precisely control their victims. To combat AI-generated scams, society must unite and establish robust defenses through technological measures, educational outreach, and emotional support. Short video platforms, in particular, must employ multi-layered techniques to identify and eliminate fraudulent accounts and detect fake celebrity videos at their source.

#### 1. Detecting Anomalous Devices  
**Dingxiang Device Fingerprinting** can distinguish legitimate users from potential fraudsters by tracking and comparing device fingerprints. This technology uniquely identifies and analyzes devices to detect malicious tools like virtual machines, proxy servers, and emulators. It also flags unusual behaviors such as multi-account logins, frequent IP address changes, and inconsistent device attributes. These insights help trace and identify fraudulent activity.  

#### 2. Detecting Anomalous Account Activity  
Anomalous activities like remote logins, device changes, phone number updates, and sudden activation of dormant accounts require frequent validation. Persistent identity verification during sessions is crucial to ensure consistent user identity.  
**Dingxiang atbCAPTCHA** can accurately distinguish between humans and bots, effectively identifying fraud in real-time. It continuously monitors and blocks anomalous behavior.  
![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)
#### 3. Preventing Deepfake Videos  
**Dingxiang Full-Chain Panoramic Facial Security Threat Perception Solution** leverages multi-dimensional data, including device environment, facial information, image forensics, user behavior, and interaction states, for intelligent verification. It quickly identifies over 30 types of malicious activities, such as injection attacks, liveness forgery, image manipulation, camera hijacking, debugging risks, memory tampering, root/jailbreak actions, malicious ROMs, and emulators.  
When detecting forged videos, fake facial images, or suspicious interactions, the system can automatically block operations. Additionally, it allows flexible configuration of video verification strength to maintain a dynamic balance between user-friendliness and security. Normal users are verified via **atbCAPTCHA**, while suspicious users undergo enhanced scrutiny.  

#### 4. Uncovering Potential Fraud Threats  
**Dingxiang Dinsight** assists businesses with risk assessments, fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight's average processing speed for daily risk control strategies is under 100 milliseconds and supports configurable multi-source data integration.  
With mature indicators, strategies, and deep learning models, Dinsight enables self-monitoring and iterative risk control mechanisms. Paired with the **Xintell Intelligent Model Platform**, it optimizes security strategies for known risks, mines potential risks through log analysis and data exploration, and applies one-click configuration for various scenarios. Based on correlation networks and deep learning, Xintell standardizes data processing, feature engineering, model development, and deployment into a one-stop modeling solution.  

By combining advanced technology, proactive education, and emotional support, platforms can significantly reduce the risks posed by AI-driven fraud, ensuring safer online experiences for all users.

Elderly Woman Scammed Out of 2 Million Yuan by AI-Generated "Celebrity

Each year during the Double 11 shopping festival, major e-commerce platforms roll out various promotional strategies, such as coupons, tiered discounts, and pre-sale deposits, to attract consumer

Each year during the Double 11 shopping festival, major e-commerce platforms roll out various promotional strategies, such as coupons, tiered discounts, and pre-sale deposits, to attract consumer participation. While some shoppers manage to purchase their desired items through careful planning, many encounter issues like fake discounts, price manipulation, and subpar product quality, which result in consumer rights violations. These disputes not only harm user experience but also hinder the healthy development of the e-commerce market.

# Large-Scale "Inflated-then-Discounted" Pricing

Some consumers have reported noticeable "inflated-then-discounted" price fluctuations on products during promotions, with some prices even increasing after deposits were paid.

![2024111403.jpg](https://cdn.aisecurius.com/images/dca/dcaad7cc-d281-46d4-ae57-5f23a6d492d8.jpg)

Pan Lei, a consumer, planned to purchase two down jackets originally priced at 2,499 yuan each, with a promotional final price of 1,644.01 yuan per item. After paying a 250-yuan deposit, the remaining balance due on October 21 was supposed to be 1,394.01 yuan. However, when it came time to pay, the balance had risen to 1,706.05 yuan, making the total cost over 200 yuan higher than expected. After unsuccessful negotiations, Pan opted to return the items before the deadline and filed a complaint with the 12315 Consumer Rights Protection Platform.

Similar cases occurred with other products. One consumer found the balance on an item, initially quoted at 149.46 yuan, had risen to 195.86 yuan. Despite keeping screenshots of the initial price, customer service did not provide a satisfactory response. Another consumer noticed the price of a jacket in their shopping cart increased from 296.1 yuan to 329 yuan, while a silk quilt rose from 309 yuan to 389 yuan, with final promotional prices even higher than before the sale period.

Such frequent price changes have eroded consumer trust in promotional activities. For example, Hong Ming purchased two sets of books for 188.2 yuan during the event, only to find the same books priced at 177 yuan post-sale. Customer service explained the price fluctuation was due to varying discounts and compensated him with an 11-yuan voucher, which Hong found insincere.

Price manipulation is not new on e-commerce platforms. Despite increased regulatory efforts, some merchants still engage in deceptive practices, undermining consumer experiences during Double 11 and casting doubt on the platforms' integrity.

Consumers have also reported receiving items that did not match their orders or were misrepresented due to false advertising. Some merchants delayed shipping for extended periods, often providing misleading reasons and pushing consumers to cancel and reorder at higher prices. Such behaviors represent classic cases of consumer fraud, infringing on legitimate rights and disrupting the fair market environment.


# Ambiguity in Customer Service Responses


Regarding merchants’ "inflated-then-discounted" pricing practices during Double 11, one e-commerce platform stated it would monitor promotional product prices and penalize merchants engaging in malicious price hikes through warnings, deductions, or removal from the platform. Another platform’s customer service attributed balance changes to the expiration of promotions or price adjustments, advising consumers to verify the reasons for price variations.

![2024111401.jpeg](https://cdn.aisecurius.com/images/4ac/4acafb2a-7e8a-4269-853d-8163819dbda6.jpeg)

To regulate promotional practices, the State Administration for Market Regulation issued the *Compliance Guidelines for Double 11 Online Promotions*, mandating platforms to optimize promotional rules, prohibit price fraud, and ensure transparency. Experts noted that tactics like price inflation before discounts or increased final payments constitute price fraud, misleading consumers through false pricing. According to the *National Development and Reform Commission’s Provisions on Prohibiting Price Fraud*, merchants must display accurate pricing, ensure promotional prices are lower than original prices, and provide evidence for original prices without arbitrary markings.


# Recommendations for Market Order and Consumer Protection


The Dingxiang Defense Cloud Business Security Intelligence Center suggests strict regulatory action against false promotions, substandard products, and other issues to ensure merchants fulfill their commitments. Consumers should choose reputable sellers, engage in rational consumption, and prioritize after-sales protection over low prices.

![2024111402.png](https://cdn.aisecurius.com/images/1a8/1a8c0885-21a3-438b-b83c-ed00de72399a.png)

When encountering disputes during online shopping, the Dingxiang Defense Cloud Business Security Intelligence Center suggests several approaches for consumers to safeguard their rights effectively:

1. **Direct Negotiation with the Merchant**  
   Consumers should first try to directly communicate with the merchant, explain the issue, and clearly state their demands. Many disputes can be resolved through open dialogue.

2. **Seek Assistance from the Platform**  
   If negotiations fail, consumers can contact the platform's customer service or use the complaint channels provided by the e-commerce platform to request intervention. Platforms typically have rules and service guarantees in place to mediate disputes and provide fair resolutions.

3. **Involve Consumer Rights Organizations**  
   If the issue remains unresolved after platform intervention, consumers can report their cases to consumer rights protection associations or relevant industry organizations. These groups can pressure merchants to fulfill their obligations and offer legal and regulatory guidance.

4. **Pursue Legal Action**  
   As a last resort, consumers can file a lawsuit to assert their legitimate rights. Although this option may be more complex, it is an effective method when sufficient evidence is available. By following these steps, consumers can choose the most appropriate means to resolve their online shopping disputes.

# Evidence Consumers Should Retain

The Dingxiang Defense Cloud Business Security Intelligence Center advises consumers to proactively collect and preserve key information and evidence, such as merchant details, product condition, transaction records, and quality verification materials, to ensure their rights are effectively protected.

1. **Merchant Information**  
   Consumers should keep their account registration and real-name authentication details to verify their identity as a party in the transaction. For merchants, it is essential to obtain information such as their name and address. According to the *E-commerce Law*, platforms are obligated to disclose necessary merchant details, which are usually accessible on major platforms.

2. **Product Information**  
   Upon receiving the product, especially for high-value items, consumers should record an unboxing video to document the product’s initial condition. Such videos serve as evidence of the product's state and can resolve disputes related to damage or quality issues. Unboxing videos capture the product’s appearance, functionality, and packaging integrity, aiding in resolving disputes caused by transportation or defects.

3. **Order and Communication Records**  
   Transaction records are critical for defending consumer rights. Consumers should save order snapshots to document pricing, payment times, and promotional discounts. Additionally, preserving communication logs, such as chat messages and call records with merchants, is vital. These records reflect the promises, negotiations, and dispute resolutions made during the purchase process, substantiating the authenticity and completeness of the consumer’s claims.

4. **Proof of Product Issues**  
   If a product is faulty, consumers should retain relevant evidence, such as inspection reports, defect certifications, photos, and videos. These materials are crucial for proving the product's issues, ensuring authoritative handling of disputes.

How Many “Inflated-then-Discounted” Products Did You Buy?

Below are six common traps consumers encountered during this year’s Double 11. These dishonest promotional practices not only harm consumer rights but also hinder the healthy development 

As the Double 11 shopping festival concluded, e-commerce platforms touted optimized shopping rules and simplified consumer experiences. However, some merchants and platforms continued to employ tactics that left consumers feeling deceived. Common practices included raising prices before the event only to lower them later, masking real discounts, and misleading consumers into thinking they had secured bargains, only to discover otherwise.

![2024111406.jpeg](https://cdn.aisecurius.com/images/cdf/cdf71228-9165-4765-bd68-a1c92a2d7e11.jpeg)

Below are six common traps consumers encountered during this year’s Double 11. These dishonest promotional practices not only harm consumer rights but also hinder the healthy development of the e-commerce market.


# Six Traps Consumers Fell Into


1. **Price hikes before discounts, misleading promotions**  
   Many consumers discovered that prices had been artificially raised prior to the Double 11 sales event. Post-discount prices were sometimes even higher than regular prices. For example, Ms. Wang from Tianjin noticed that the infant formula she regularly purchased was more expensive during Double 11 by several dozen yuan, despite displaying "savings" that seemed exaggerated. This raised questions about the merchants' pricing strategies.

![2024111404.png](https://cdn.aisecurius.com/images/6de/6dec591f-93ed-4289-b993-737e8baf59b7.png)

2. **Presale schemes lead to higher costs**  
   While presale campaigns promised savings, many consumers reported paying more in the end. After placing a deposit for a presale product, they discovered the final price after the sale was higher than buying the item outright. Merchants often refused to honor price guarantees, wasting consumers’ time and money.

3. **Misleading "lowest price" claims in live-streaming sales**  
   Live-streaming has become a popular sales channel, but some hosts falsely advertised “the lowest price.” Merchants sometimes offer exclusive discounts in live-streams, creating the illusion of a bargain, while the same product is available at a lower price on other platforms. Consumers are often misled into thinking they’ve secured a great deal.

4. **Overly complex discounts and minimal savings**  
   Platforms offered numerous discounts, including coupons, red envelopes, and tiered reductions, during Double 11. However, many consumers found the rules confusing and the savings negligible. After spending significant time participating in these activities, they discovered the discounts were less effective than simply purchasing the item outright.

5. **Poor-quality goods in live-streaming sales**  
   Despite the rapid rise of live-streaming commerce, many products sold in these sessions were of subpar quality. Top influencers often used exaggerated claims, fake pricing tactics, or emotional marketing to lure consumers. These sales channels sometimes featured counterfeit or inferior goods, exposing consumers to considerable risks.

6. **Risks of accidental purchases with password-free payments**  
   The promotion of “buy now, pay later” or “zero-cost ordering” payment options led to inadvertent purchases. The simplicity of these systems caused consumers to buy items they didn’t need or forget to make payments, leading to credit issues or other application risks.


#  Why Do Merchants Design "Price Traps?


According to the Dingxiang Defense Cloud Business Security Intelligence Center, there are several key reasons for this practice:

**First**, in the fierce battle for online traffic, advertising placements and traffic purchases on e-commerce platforms have become common ways for merchants to gain exposure. To avoid losses, merchants often resort to raising prices. This creates a dilemma: they need to attract consumers while maintaining profitability. In such scenarios, tactics like “price manipulation” and “quality misrepresentation” emerge as “necessary means” to meet promotional and platform requirements. By inflating prices first and then reducing them under the guise of promotions, merchants not only align with platform campaigns but also lure consumers with deceptive discounts, securing higher profits.


**Second**, as the popularity of the Double 11 shopping festival has gradually waned, merchants and platforms face market saturation and consumer fatigue. In response, they have started employing various price manipulation tactics, such as fake discounts, deposit-based coupons, and tiered reductions. These strategies aim to entice consumers, sometimes even misleading them into thinking they’re securing great bargains. While these tactics may attract short-term attention, they often leave consumers unknowingly spending more than anticipated.

**However**, as the Double 11 festival matures, consumers have become increasingly vigilant. This heightened awareness forces merchants to employ more “creative operations within the rules.” Yet, many of these strategies are now easily spotted by consumers.

**Finally**, while some merchants may achieve temporary traffic and sales boosts through “price manipulation,” the long-term outcomes are far less optimistic. When consumers realize they have been deceived, it often results in disappointment or anger after their purchases. This, in turn, severely damages the merchants’ credibility and brand reputation. Overreliance on such strategies undermines long-term market competitiveness and erodes consumer trust.

Six "Shopping Traps" Consumers Encountered During Double 11 Online Shopping

Recently, the Nanjing police launched a comprehensive crackdown on ticket scalping around the Nanjing Museum, successfully arresting 67 suspects and uncovering two criminal gangs that 

Recently, the Nanjing police launched a comprehensive crackdown on ticket scalping around the Nanjing Museum, successfully arresting 67 suspects and uncovering two criminal gangs that illegally used technological means to seize and resell tickets for cultural and tourist venues. 

The Nanjing Museum has a reception limit of 22,000 visitors per day, but the number of reservations often exceeds one million. Due to the scarcity of tickets, scalpers have been reselling free tickets at inflated prices. In addition to the Nanjing Museum, this also involves the National Museum, the Great Hall of the People, the Shanghai Science and Technology Museum, and more than a dozen other scenic spots. Tourists have reported that on some idle social platforms, anonymous users resell Nanjing Museum tickets, either by directly inflating the prices or bundling them with "guides" to sell them at inflated prices of up to 100-200 yuan. Behind the "hard-to-get" tickets, a gray industry chain has quietly taken shape on the internet, with clear division of labor.

![2024102901.jpeg](https://cdn.aisecurius.com/images/c61/c61a8d18-1319-412d-996d-e8d1d66a04ec.jpeg)

The police investigation found that multiple criminal gangs were behind the ticket scalping, and the division of labor within these gangs enabled rapid expansion, forming a stable profit chain. 

For example, one suspect developed "ticket-snatching software" for the Nanjing Museum and other popular cultural and tourist venues, specifically designed to quickly grab tickets. Another suspect was responsible for posting information on social media and various platforms to attract large-scale "ticket scalpers" and find target customers. Once the tickets were sold, the two split the earnings according to a set ratio. Within just over a month, the two earned nearly 300,000 yuan. 

The gang's illegal actions seriously undermined the fairness of the online reservation system for the Nanjing Museum and other cultural and tourist venues, constituting the crime of damaging computer information systems. The police have taken criminal measures against all the suspects involved to maintain market order.


# Analysis of the "Ticket-Snatching Software" Used by Scalpers


The criminal gang's "ticket-snatching software" operates with a high level of automation, completing the entire ticket purchase process in just a few seconds.

![2024102902.jpeg](https://cdn.aisecurius.com/images/82e/82eb288b-eafb-4b59-85ce-92ca2d4df0f8.jpeg)

High-Frequency Access: The software uses multiple IP addresses from different areas in Guizhou to make high-frequency requests, sending a large number of requests in a short period, creating an abnormal situation that human operators cannot achieve.

Millisecond-Level Request Intervals: The time intervals between requests are compressed to milliseconds, with the system receiving more than 2,000 visits per day. This behavior can be metaphorically compared to "thousands of people with the same face and clothes, bypassing the community security gate, and standing at your door, coming thousands of times a day."

Speed is crucial in ticket-snatching. If person A is one second faster than person B, A will be able to purchase the ticket while B will not. In terms of speed, humans rely on their neurological reactions, but software operates based on pre-set processes, making it much faster than humans, significantly increasing the success rate of ticket purchases.


# Preventing Scalpers from Using Software to Snatch Tickets


Although many museums have implemented ticketing rules and technical measures to regulate scalpers, such as blocking frequent reservations from specific phone numbers or IP addresses, scalpers continue to develop new cheating tools. These ticket-snatching tools can perform batch registration, login, and ticket grabbing operations, enabling them to quickly and automatically seize designated tickets in an instant. By filling in personal information, setting the quantity, and scheduling the running time, they can complete the automatic purchase process.

Therefore, simply limiting IP addresses is not enough and may inadvertently block legitimate visitors' reservations. Dingxiang Defense Cloud Business Security Intelligence Center recommends that, in addition to restricting IP addresses and accounts, more technical measures should be implemented to prevent scalpers from using cheating software to effectively curb the reselling of tickets.


## Ticketing Rule Restrictions


1. **Increase the usage limit of identification documents.** Each identification number can only make one reservation; each account can reserve tickets for up to 5 people.

2. **Increase account usage limits.** Some museums require that if an account makes multiple reservations within a week and has a no-show rate exceeding 50%, that account (or phone number) will be restricted from making reservations for 30 days. Some museums also require that if an account cancels reservations more than three times within 7 days or five times within 30 days, or fails to enter the venue after making a reservation, the account will be locked in a "blacklist" for 30 days and cannot make any reservations, to prevent scalpers from repeatedly using identity information they control.

3. **Increase ticket sale times and channels.** Irregular ticket releases and the addition of manually issued random tickets can help reduce the occurrence of technical ticket-snatching to some extent.


## Technical Measures for Prevention


1. **Detect and identify abnormal devices.** Identify whether the client’s Device Fingerprinting is legitimate, and check for risks such as injection, hooks, emulators, and more. Quickly identify risks related to flashing, rooting, jailbreaking, hijacking injections, and abnormal behaviors such as multiple activations on the same device, abnormal IP associations from the same device, abnormal behavior patterns in older device models, and abnormal use of old operating systems within the same channel.

Dingxiang Device Fingerprinting can detect virtual machines, proxy servers, emulators, and devices that are maliciously controlled. It analyzes whether devices exhibit abnormal or unusual behaviors, such as frequent switching of accounts, frequent IP address changes, or frequent changes in device attributes. This helps track and identify fraudulent activities early on and prevent fraudulent claims. It can also serve as an additional identity verification factor, enhancing the security of user logins and transactions. By recording and comparing Device Fingerprints, legitimate users and potential fraudulent activities can be distinguished.

2. **Detect and intercept abnormal accounts.** Based on user behavior, strategies can be deployed to monitor accounts that initiate orders after switching a large number of accounts on the same device.

Dingxiang atbCAPTCHA can perform real-time verification, identification, and interception of malicious accounts and crawling behaviors during key actions like registration, login, and queries. Based on AIGC technology, it can prevent AI-based brute force cracking, automated attacks, and phishing attacks, effectively preventing unauthorized access and blocking web crawlers from stealing data. It integrates 13 verification methods and multiple preventive strategies, supports seamless user experience for legitimate users, and reduces real-time response time to under 60 seconds, further enhancing the convenience and efficiency of login services.

3. **Enhance risk identification and prevention capabilities.** Establish a local dynamic operation and maintenance mechanism for black and white lists based on registration data, login data, and activation data, including user IDs, phone numbers, devices, and other dimensions. After accumulating a certain amount of online data, the risk control data and business accumulation can be used to model behaviors like registration, login, ordering, and ticket-snatching. The model’s output can be directly applied to risk control strategies.

Dingxiang Dinsight helps businesses perform risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. The Dinsight system, along with the Xintell intelligent model platform, can automatically optimize security strategies for known risks, mine potential risks based on logs and data, and configure one-click risk control strategies for different scenarios.

**The police remind tourists** not to trust false ticketing information regarding "ticket-snatching services" on social media platforms. These services pose risks of personal information leakage and may even lead to scams. If you encounter any suspicious activity, please call 110 to report it and protect your rights.

Nanjing Arrests 67 "Scalpers" Suspected of Reselling Museum Tickets

Ms. Li was surprised to see her full name, address, and phone number on the package, including a real name she rarely uses for online shopping. 

Recently, consumers across various regions in China have reported receiving suspicious packages labeled as "prize winnings." These packages were addressed accurately with recipients’ personal details and often contained a scratch card or flyer promoting valuable prizes. Unfortunately, some individuals fell for these scams, losing up to 45,000 yuan. It is critical not to trust such "prize packages."


# Two Notable "Winning Package" Scam Cases



### Case 1: Ms. Li from Chengdu

Ms. Li, a resident of Chengdu’s Wuhou District, received a message on October 28 about a Yunda Express package ready for pickup. Believing it to be an online order, she collected it only to find an envelope with a thank-you letter and scratch card inside. The card revealed a prize worth “thousands of yuan,” including a gift set, cash bonuses, and more. However, to claim these prizes, she was instructed to scan a QR code or contact customer service, providing her personal information.

![2024110801.png](https://cdn.aisecurius.com/images/64f/64fef792-730c-4292-9e38-803028fa2122.png)

Ms. Li was surprised to see her full name, address, and phone number on the package, including a real name she rarely uses for online shopping. Growing suspicious, she checked online and found other users reporting similar experiences. Some victims had scanned the code, provided personal details, and fallen for the scam.


### Case 2: Ms. Liu from Jiangsu

Another case involved Ms. Liu from Jiangsu, who also received a "winning card" and scanned the QR code, joining a "gift redemption" group chat. In the group, a customer service representative instructed her to download a specific app, listen to music, and share screenshots for small cash rewards. After a few successful earnings, Ms. Liu felt reassured about the group’s legitimacy. She was later encouraged to participate in a "competitive task," with promises of higher returns for larger investments. Feeling tempted, she invested 45,000 yuan but was told there was an "error" and asked to transfer an additional 45,000 yuan. Realizing she’d been scammed, she reported the incident to the police, who have since begun an investigation.
![2024110802.png](https://cdn.aisecurius.com/images/030/0305c6c2-d4f1-477f-b886-56b9e79a0809.png)
Statement from Targeted Brands
Responding to consumer feedback and complaints, a well-known snack brand that was impersonated in these scams, issued a statement clarifying that they had not created, distributed, or authorized any such cards or promotions. They urged consumers not to scan any unknown QR codes or share personal details and to verify information sources carefully.


# Characteristics of These Scams


**Accurate Data Access:** Scammers possess detailed information, including the recipient’s name, address, and phone number, lending legitimacy and making it harder for victims to doubt the package’s authenticity.

**Combined Online-Offline Approach**: The scam utilizes the physical trust associated with packages, combined with online processes like QR codes and customer service contacts, creating a multi-step deception.

**Layered Manipulation**: Victims are led through multiple steps—from joining group chats to downloading unauthorized apps—to build trust gradually, making them more susceptible to subsequent scams.

**High-Quality Imitation**: Scammers exploit brand names and distribute small cash prizes or physical items to reinforce credibility, convincing victims to continue participating.

Consumer Precautions

The Dingxiang Defense Cloud Business Security Intelligence Center suggests that consumers enhance their vigilance with the following measures:


### Do Not Trust Unknown Packages Easily
: When receiving a package from an unknown source, maintain caution. For unfamiliar prize notifications—especially if they involve scanning a QR code or contacting unknown customer service—avoid providing personal information.


### Be Skeptical of Various Rewards
: Scammers often use small cash rewards or gifts to lower the victim’s guard, then gradually increase the stakes. Remember that “nothing comes for free,” and maintain a healthy level of suspicion towards any unusually high returns.


### Avoid Downloading Apps from Unverified Sources
: Be particularly wary of apps that cannot be found on mainstream platforms or official app stores. Downloading such apps may lead to personal information leakage or even financial risk.


### Verify and Seek Help from Authorities in Doubtful Situations
: If you suspect a potential scam, check online or contact the local police hotline for information. When in doubt, report it to the police to prevent financial loss.

In conclusion, these “winning package” scams leverage couriers, online chat groups, and other forms of allure to create a high level of deception. Consumers should enhance their awareness, carefully verify the source of information, avoid trusting prize messages easily, and protect themselves from fraud.



Several "Winning Package" Scams Reported During This Year's Double 11 Shopping Festival

During the Double 11 Shopping Festival, while consumers enjoy shopping, they also face potential fraud risks. In live-streaming rooms

During the Double 11 Shopping Festival, while consumers enjoy shopping, they also face potential fraud risks. In live-streaming rooms, screens are often flooded with positive reviews and likes that can easily attract buyers. However, when consumers place orders based on high sales and five-star reviews, they often find that the actual products fall far short of expectations. Behind this phenomenon are often various "Internet Water Army" groups that create fake orders and control reviews to fabricate false sales data and positive reviews for merchants.
![2024102401.png](https://cdn.aisecurius.com/images/e09/e0996a44-4ba2-4663-bdfa-b018bc04d6ce.png)
Recently, the Neijiang Police in Sichuan Province uncovered a major case involving a "network water army." Over the past two years, this gang provided fake order and review control services to over 1,000 merchants on multiple online platforms, with a total involvement of 100 million yuan. This group not only boosted “popularity” in live-stream rooms but also participated in fake interactive comments and used “empty packages” to create fake orders.

This behavior not only disrupts normal market order but also severely affects consumers' shopping experience.


# Both an Internet Water Army and an E-Commerce Fake Order Scheme


The police investigation revealed that this “network water army” gang targeted live streaming and offered three types of services: Rank Boosting, Interaction, and Marketplace Exposure.
![2024102402.png](https://cdn.aisecurius.com/images/7e2/7e27efaf-1cf4-42ec-89cb-18e7b1ffc8f0.png)

- **Rank Boosting**: The water army increases the number of viewers in a live-streaming room by appearing as named viewers, directly affecting the room's popularity without interacting. The primary purpose is to boost superficial popularity.
  
- **Interaction**: Besides rank boosting, the water army also participates in interactions within live streams, including commenting and answering viewer questions. Merchants provide discussion topics, and sometimes the water army improvises to create a realistic interactive atmosphere. This service charges 4 yuan per person.

- **Marketplace Exposure**: The water army enters live-stream rooms via the marketplace page, further increasing the product's exposure rate. This service charges 5 yuan per person, helping merchants attract more potential customers.

Apart from acting as an “internet water army,” this gang also engages in advanced fraud schemes involving fake orders. Live-stream merchants use the water army to buy specified products to boost sales data, then ask them to report the order numbers. By mailing “empty packages,” they create fake orders. After receiving the empty package, the water army gives five-star reviews and false comments as instructed, earning commissions based on the product price.

![2024102403.png](https://cdn.aisecurius.com/images/cf3/cf30a3b3-8f79-4482-9fbe-c736c55df702.png)

- **Fake Ordering**: Merchants instruct the water army to purchase specific products to inflate sales figures. The water army then provides the order numbers back to the merchant.

- **Sending Empty Packages**: Merchants send empty packages to the water army, who does not receive actual products but instead follows instructions to leave five-star reviews and comments. This effectively creates fake orders and builds a false impression of high sales.

- **Commission Settlement**: The water army earns commissions based on product prices, incentivizing continued participation in fake order activities. Commission payments are typically tied to the purchased product price.

New members joining this gang are required to pay a membership fee ranging from 199 yuan to 2,999 yuan. Depending on the fee amount, members receive titles such as Founder, Director, or Trainee and are assigned different roles like sales, shills, or customer support. Since March 2022, the gang has manipulated reviews and fake orders for more than 1,000 merchants across multiple platforms. By recruiting new members and charging referral fees, they have illegally profited from their schemes, with a total amount involved reaching 100 million yuan. The main suspects are currently under criminal detention in accordance with the law.


# Technical Measures to Prevent Water Army and Fake Orders


This criminal group reached a scale of over 100,000 people, with each member account operated by a real person and different IP addresses, making it difficult for regulatory authorities to identify "network water armies." Analysis reveals that water army accounts may exhibit high activity shortly after registration or at specific times, while showing little to no activity at other times. Additionally, water army accounts often have lower user levels and credit scores, which differ significantly from those of real users in terms of level and permissions.

To effectively curb these illegal activities, platforms need to leverage technical means to identify and prevent suspicious behaviors, improving their ability to detect water army accounts and suspicious activities.

---

**1. Identifying Abnormal IP Addresses and GPS Locations**

By monitoring IP addresses and GPS information, platforms can detect unusual login behavior. For example, if an account is accessed from multiple geographic locations in a short time, or if the IP address does not match the registered information, the system can flag the account as suspicious. Additionally, IP address whitelist and blacklist management can effectively filter known sources of water army activity, reducing potential risks.

**2. Detecting Client Device Fingerprinting**

Detecting the legitimacy of client device fingerprinting is another important method to prevent water armies. By collecting data on device operating system versions, hardware information, user behavior, and CPU instructions, the platform can identify risks like emulators, hijacking, or injection. Emulators typically cannot fully replicate carrier information and other system characteristics, making it easier to detect cheating.

Dingxiang Device Fingerprinting records and compares fingerprints to distinguish legitimate users from potential fraudsters. It assigns a unique identifier to each device, detecting maliciously manipulated devices, such as virtual machines, proxy servers, and emulators. It analyzes devices for multiple account logins, frequent IP address changes, and modifications to device properties, helping track and identify fraudulent activities.

**3. Monitoring Abnormal User Behavior and Patterns**

Once a platform has accumulated sufficient online data, it can model user registration, login, order, and purchasing behaviors through risk control and operational data. Using machine learning and data analysis, unusual behavior patterns—such as frequent registration or ordering—can be identified. The output of these models provides effective support for risk control strategies, allowing for timely detection and response to abnormal activities.

**4. Identifying Abnormal Accounts**

Quickly identifying suspicious accounts is key to preventing fake orders. Platforms can monitor several types of suspicious behaviors, such as repeated activations on the same device, associations between one device and multiple IP addresses, a high concentration of logins from the same IP in a short time, and unusual ratios of old device models or operating systems within a particular channel. Maintaining a dynamic whitelist and blacklist can help the platform continuously update and manage data, such as user IDs, phone numbers, and device information, blocking suspicious accounts in a timely manner.

Dingxiang Dinsight assists businesses with risk assessment, fraud detection, and real-time monitoring to enhance the efficiency and accuracy of risk control. Dinsight’s average processing speed for daily risk control strategies is under 100 milliseconds. It supports multi-source data configuration and integration, using well-developed indicators, strategies, and model experience, as well as deep learning techniques. This enables real-time self-monitoring and self-iteration. In conjunction with Xintell’s intelligent modeling platform, known risks can be automatically optimized. Based on risk control logs and data mining, potential risks can be identified and configured with a single click to support risk control strategies in various scenarios. With deep learning technology, Xintell standardizes complex processes, providing one-stop modeling services from data processing and feature engineering to model building and deployment.

**5. Strengthening Mechanisms to Block Abnormal Accounts**

To improve prevention effectiveness, platforms need to intercept abnormal accounts by increasing the difficulty of verification elements, such as frequently updating the verification image library to raise the operational cost for water armies and fake order participants. Additionally, comparing verification environment information with normal user behavior helps quickly detect anomalies, further securing the platform.

Dingxiang atbCAPTCHA, based on AIGC technology, can prevent threats from AI brute-force cracking, automated attacks, and phishing, effectively blocking unauthorized access, account theft, and malicious operations, thus protecting system stability. It integrates 13 verification methods and multiple risk control strategies, combining 4,380 risk policies, 112 categories of threat intelligence, and covers 24 industries with 118 types of risks. Its detection accuracy reaches 99.9%, and it quickly transforms threats into actionable intelligence. With real-time response capabilities reduced to 60 seconds, it also enhances the convenience and efficiency of digital login services for genuine users.


This Fraud Ring Has Been Caught! Mainly Boosted Online Influencers by Fake Orders and Assembled Internet Water Armies

Mr. Wang, a citizen, received a FaceTime call displaying the name “China Banking Regulatory Commission.” The caller, posing as a representative, informed him that the government had lowered loan 

Mr. Wang, a citizen, received a FaceTime call displaying the name “China Banking Regulatory Commission.” The caller, posing as a representative, informed him that the government had lowered loan interest rates and asked if he had any loans, offering to help lower his rate. Since Mr. Wang did indeed have a mortgage and had also used a consumer loan platform, he believed the caller. He agreed to the caller’s screen-sharing request. The caller also provided a website link, which appeared to be an official site. Following the “representative’s” step-by-step instructions, Mr. Wang not only transferred tens of thousands of yuan from his account but also took out two or three loans from lending platforms, suffering a total loss of over 90,000 yuan.

![2024110101.png](https://cdn.aisecurius.com/images/7ae/7ae2c334-e3c4-4d0b-bd9a-6ae57db81926.png)

Recently, reports have surged nationwide of scams involving iPhone FaceTime calls. These new fraud tactics are constantly evolving, even incorporating “AI face-swapping” technology to deceive victims. Police advise verifying the authenticity of suspicious calls.


# FaceTime Screen-Sharing Scam Process


FaceTime, an in-built feature on Apple devices, offers a screen-sharing function, and all reported victims are iPhone users. During a FaceTime call, the scammer only needs to send a screen-sharing request, and once the victim accepts, all actions on the victim’s screen become visible to the scammer. The scammer can then monitor the victim’s every move, watching them input sensitive information such as bank account numbers and passwords.
![2024110102.jpg](https://cdn.aisecurius.com/images/bc1/bc1b5da3-b4b4-415b-88fd-d4310ddfb6c3.jpg)

1. **Establishing Trust with the Victim**  
   Scammers use FaceTime’s ability to change the display name, posing as reputable institutions like the “China Banking Regulatory Commission,” “China UnionPay Service,” “Alipay Customer Support,” or even as law enforcement agencies to deceive the victim. Most users, seeing such “trustworthy” caller ID, lower their guard and become more susceptible to fraud.

2. **Exploiting Screen-Sharing to Steal Sensitive Information**  
   FaceTime’s screen-sharing function allows the scammer to see the victim’s screen in real time, guiding the victim to input sensitive information like login details and passwords. The scammer monitors the screen, capturing this valuable data without the victim’s awareness.

3. **Guiding the Victim’s Actions**  
   The scammer may coax the victim to make a transfer or direct them to a fake official website or phishing link, ultimately prompting the victim to enter banking details, passwords, or other sensitive data. If the victim hesitates, the scammer may intercept a verification code to execute the transaction independently, transferring funds without the victim’s knowledge.


# Preventive Measures for Apple iPhone Users


1. **Verify Identity Through Multiple Channels**  
   Law enforcement, banks, and government agencies typically do not use FaceTime or similar video apps to contact users. If in doubt, verify the identity through officially published contact information rather than trusting the displayed caller name.

2. **Disable the FaceTime Sharing Feature**  
   If not needed for daily use, go to “Settings” on the iPhone, find the “FaceTime” option, and turn it off to prevent scammers from initiating fraud through FaceTime.

3. **Avoid Screen Sharing with Strangers**  
   Never share your screen with strangers, regardless of the video software used. Be especially cautious if the caller claims to be an official representative and requests to view personal account activities.

4. **Be Wary of Phishing Websites and Fake Apps**  
   Do not click on unfamiliar links, especially those provided during screen sharing. Only download apps from the Apple Store to avoid malicious links or fake applications.

5. **Protect Personal Information and Privacy**  
   Do not disclose personal sensitive information such as facial recognition data, bank card numbers, dynamic verification codes, or account passwords to strangers. Maintain a high awareness of information security to prevent exploitation.

6. **End Suspicious Calls Promptly**  
   If you receive a suspicious video call or call, stay calm, make an excuse about poor signal, hang up, and immediately verify through another method. Avoid responding directly to potential scams.

In addition, Dingxiang's device fingerprinting technology can detect screen-sharing calls from apps in real time and issue immediate alerts to remind users to stay secure. When screen sharing is activated, Dingxiang’s device fingerprinting detects this and notifies users of potential risk.

FaceTime screen-sharing fraud is a new type of scam that uses video calls and screen-sharing features to deceive victims into unknowingly exposing account and personal information. By understanding the characteristics of such scams and enhancing security awareness, users can better protect themselves from potential losses. Increasing awareness of video sharing risks and maintaining vigilance against phishing sites can effectively reduce the likelihood of being scammed.

Apple iPhone Users Alert! Scammers are Using FaceTime for Fraud

Ms. Que from Chengdu encountered a typical online scam. She received a delivery package with a card reading “Claim Three Free Gifts for ¥0.”

As the Double 11 shopping festival kicks off, while consumers are busy purchasing favorite items, criminals are taking advantage by setting up various online fraud schemes. The Cybersecurity Bureau of the Ministry of Public Security advises the public to stay alert to avoid falling into these scams.

![2024102202.png](https://cdn.aisecurius.com/images/e3e/e3e1b706-fe26-4c5f-aa25-03552a9ef73f.png)

1. **Impersonated Customer Service Scams**: Scammers may pose as customer service representatives, sending fake messages or emails to request personal information or payments. Consumers should verify the customer service identity through official channels and avoid sharing personal details readily.

2. **Fake Shopping Websites**: These sites appear legitimate and attract consumers with low prices, but they’re traps set by scammers. Consumers should only shop on well-known e-commerce platforms and avoid direct payments to unknown sites.

3. **Pre-sale Scams**: Taking advantage of Double 11 pre-sale events, scammers deceive consumers into paying upfront, only to disappear afterward. Consumers should stick to pre-sale activities with platform guarantees and avoid paying strangers directly.

4. **Prize-Winning Scams**: Scammers may notify consumers by text or email about “winning” a prize, requiring a fee or taxes to claim it. Consumers should remember that legitimate prize draws do not ask for fees upfront.

5. **Delivery Scams**: Scammers posing as couriers send fake messages, tricking consumers into following a public account to steal their information. When receiving such messages, consumers should contact the delivery company directly for verification.

6. **Return Scams**: Using returns and refunds as an excuse, scammers trick consumers into providing one-time passcodes, gaining access to personal information or funds. Consumers should initiate returns through official channels and avoid trusting unknown calls or messages.


# Analysis of a Delivery Scam Process


Ms. Que from Chengdu encountered a typical online scam. She received a delivery package with a card reading “Claim Three Free Gifts for ¥0.” Tempted, she scanned the QR code on the card and added a supposed “customer service” account. She was then added to a group promoting cash-back schemes for small purchases and ultimately invested over ¥30,000. When she tried to withdraw her principal, the scammers gave various excuses and demanded additional deposits. Only then did she realize it was likely a scam.

These scams typically proceed in stages: they start by luring victims with free gifts or discounts, then lead them to join a cash-back group or download a fraudulent app. They establish trust through small rewards and ultimately refuse withdrawals, urging further investment. This scam tactic is highly concealed, technically sophisticated, and has a broad impact.

![2024102201.png](https://cdn.aisecurius.com/images/b64/b6485fdf-8975-4466-ae3e-3bc99738b9ec.png)

To prevent falling victim to such scams, the public should stay vigilant and avoid easily believing information or promises from strangers. Exercise caution with any financial transactions or investment opportunities, and be wary of sharing personal information, such as bank details or ID numbers. Safety tips from telecom operators and public security agencies should be carefully read and remembered.


# How Companies Can Prevent Fraudulent Deliveries


Dingxiang's Business Security Intelligence Center suggests that consumers should verify unknown deliveries by checking logistics and courier information on official websites and contacting the courier by phone. For unfamiliar packages, it’s essential to verify the courier's identity and inspect package contents. Additionally, consumers should avoid scanning QR codes or clicking links in unsolicited packages, especially if asked to provide sensitive information like bank or ID numbers.

To help logistics companies identify suspicious packages, Dingxiang recommends integrating its business security intelligence. This allows for screening high-risk numbers and staying informed on the latest fraud tactics to identify packages posted by fraud groups effectively.

Dingxiang’s Business Security Intelligence maintains over 30,000 risk sources, monitoring black-market communities, darknet forums, illegal sites and apps, crowdsourcing platforms, and legitimate data sharing to collect risk data. By profiling risk populations, scoring behavior, analyzing relationships, and detecting fraud networks, this service can provide insights into telecommunications fraud, IP and device risk, and transaction risk across various industries. Using a fraud risk verification module, it cross-references phone numbers in banking, transactions, and transfers to check if they are associated with fraud, scoring their risk levels.

Dingxiang also recommends courier companies use the Xintell Intelligent Modeling Platform to analyze data on senders, recipients, phone numbers, and devices. By leveraging Hadoop, custom components, and drag-and-drop operations, the platform simplifies big data analysis, making it accessible to engineers, data scientists, and business staff alike. Xintell supports anti-fraud efforts by providing model training for fraud detection, risk management, marketing, and customer segmentation, dynamically upgrading models over time.

Furthermore, courier companies should enhance training for couriers to help them recognize suspicious packages and handle unusual situations properly.

Effective prevention starts at the source. By monitoring for risks in shipments from the start, courier companies can significantly reduce losses and better safeguard recipients' financial security.


# Consumer E-Commerce Safety Warning


Online shopping festivals can be both exciting for consumers and profitable for fraudsters. Criminal groups often use sophisticated and concealed methods to target consumers on a large scale.

- **Verify Information**: Always verify the sender’s identity through official channels before providing personal information or making payments.
- **Choose Reputable Platforms**: Use known e-commerce platforms and avoid shopping on unknown sites.
- **Protect Personal Information**: Avoid sharing personal details, especially bank account numbers, passwords, and verification codes.
- **Be Wary of Suspicious Links**: Avoid clicking unknown links to prevent data theft.
- **Report Suspicious Activity**: Immediately report any suspected fraud to authorities.

The Double 11 shopping festival is a peak period for fraud. Consumers should take precautions to enjoy their shopping safely and protect their personal assets.

Be Extra Cautious of This Kind of Delivery During This Year’s Double 11 Shopping Festival

On September 1, Professor Li Ling from the National School of Development at Peking University issued a public statement saying, “

“I believe this not only damages my own interests but, more importantly, harms public interests, especially for seniors who prioritize health and are drawn to various health remedies. I urge elderly friends to be cautious and verify the source before making purchases. Ask yourself, is this person an expert in this field? For example, I'm a policy expert, not a medical specialist, so I would never recommend such treatments.”
![2024101501.png](https://cdn.aisecurius.com/images/c29/c29e4e75-91c3-49fa-b0e9-57b6d6eeb7cf.png)

On September 1, Professor Li Ling from the National School of Development at Peking University issued a public statement saying, “Recently, numerous fake videos have appeared online, falsely using my image to promote medical books and treatment methods. These videos are largely AI-generated. I have never endorsed any medical books or products on any platform. Please be aware of this scam.” She has since reported the matter to the police.
![2024101502.png](https://cdn.aisecurius.com/images/2e0/2e0191f5-04f5-4c84-b6a2-5ef578290191.png)

Celebrities hold significant social influence and public trust. Scammers use the reputations of well-known individuals, like Professor Li, to create trust in their fake endorsements, especially among seniors concerned with health. By cloning her likeness with AI, fraudsters present a believable endorsement, making it harder for elderly viewers to doubt its authenticity. Apart from Professor Li, famous figures like Scarlett Johansson, Kylie Jenner, Taylor Swift, and Tom Hanks have also been impersonated in similar scams.


# The Scam Process in Fake Celebrity Ads


Fraudsters rely on AI technology to create fake celebrity advertisements in a multi-step scam, from data collection and AI-generated videos to ad distribution and tricking victims into transferring money. By exploiting celebrity trust and technology, they create highly convincing scams.

1. **Data Collection**: Fraudsters begin by gathering public information on the targeted celebrity. This includes photos, videos, public speeches, interviews, and social media posts, all of which provide material for AI-generated content. For Professor Li’s case, scammers likely obtained footage of her public speeches or academic lectures to capture her voice, tone, and facial expressions.

2. **Fabrication of AI-Generated Audio and Video**: With sufficient material, scammers use AI to synthesize audio and video. AI technology can accurately replicate a person’s voice and facial expressions to create realistic fake advertisements. In Professor Li’s case, the fraudsters likely produced a video where she appears to endorse health products or medical books she has never promoted.

3. **Spreading Fake Ads**: The resulting fake video is then spread across various online platforms, such as short video apps, social media, and group chats. These platforms are frequently accessed by older adults, the scammers’ primary target audience. Using targeted marketing, the fake ads reach specific groups. Due to the striking similarity of these videos to genuine content, viewers are easily deceived.

4. **Creating a Sense of Urgency and Authority**: To boost their success rate, scammers often use scare tactics or emphasize the authority of the content. For example, they might exaggerate the risks of certain diseases or claim the promoted product is a “limited-time offer” or “exclusive.” For health-conscious seniors, fraudsters may even fake expert qualifications, fabricating scientific evidence to strengthen trust.

5. **Inducing Purchases**: When convinced, the victim is prompted to buy the fake product through links or customer service hotlines embedded in the video. In Professor Li’s case, scammers might include a link or QR code in the video, prompting victims to pay for bogus “medical books” or “treatment products.” Often, these products are either non-existent, of poor quality, or completely worthless, resulting in financial loss for the victim.


# How Ordinary People Can Identify Fake Celebrity Ads


Scammers leverage AI to create fake celebrity endorsements, following a multi-step process: data collection, synthetic audio/video production, ad distribution, and finally, persuading victims to transfer funds. How can regular people detect these fake celebrity endorsements?

1. **Be skeptical of shocking or “too good to be true” statements, especially involving celebrities**. Always apply common sense and logic.
2. **Look for unnatural features** in the content, like odd blinking patterns, mismatched audio, or distorted body parts, as these could signal deepfakes.
3. **Cross-check statements with reliable news media and websites**. Don’t believe content without fact-checking first.
4. **Before making a purchase or decision, consult friends, family, or colleagues for advice** and do some research. Don’t take the initiator’s words at face value, even if it’s someone like Elon Musk.
5. **Avoid commenting, sharing, or even clicking on unverified social media posts**. Interacting with accounts impersonating celebrities increases your scam risk.
6. **Never share personal details with unverified individuals, websites, or apps**.


# Platform Measures Against Fake Celebrity Ads


For fake celebrity ads and information on social media platforms, the Dingxiang Defense Cloud Business Security Intelligence Center suggests precise identification of fraudulent accounts as the first step in effective fraud prevention.

1. **Identify abnormal devices**. Dingxiang Device Fingerprinting helps distinguish legitimate users from potential fraudsters by recording and comparing device fingerprints. It uniquely identifies devices, detecting malicious use of virtual machines, proxy servers, emulators, etc., while analyzing behaviors like multi-account logins, frequent IP address changes, and device property changes that don’t match normal user patterns.

2. **Detect suspicious account operations**. Unusual activities like logging in from different locations, switching devices, changing phone numbers, or sudden activity on dormant accounts require stronger verification. Continuous identity verification throughout sessions is essential. Dingxiang’s “No Touch” verification can quickly and accurately distinguish human users from bots, helping real-time monitoring and blocking of abnormal behavior.

3. **Prevent face-swapped fake videos**. Dingxiang’s full-spectrum facial security threat detection solution utilizes multi-dimensional data, including device environment, facial information, image authenticity, user behavior, and interaction state for intelligent verification. It quickly detects and blocks over 30 types of malicious behavior, including injection attacks, live forgeries, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, and emulator usage. This solution flexibly adjusts video verification strength to allow seamless verification for legitimate users while increasing security for suspicious accounts.

4. **Identify potential fraud threats**. Dingxiang Dinsight aids businesses in risk assessment, anti-fraud analysis, and real-time monitoring, enhancing risk control efficiency and accuracy. Dinsight processes daily risk control strategies in less than 100 milliseconds on average, supports multi-party data integration, and leverages established indicators, strategies, and deep learning technologies for self-monitoring and iterative risk control. Coupled with the Xintell intelligent modeling platform, it automatically optimizes security strategies for known risks, configuring different scenarios to support risk control strategies. Using associative networks and deep learning, it standardizes complex data processing, feature derivation, model building, and deployment for an end-to-end modeling service.

How Ordinary People Can Identify Fake Celebrity Advertisements

The Hanyang Branch of the Wuhan Municipal Public Security Bureau received a report from a bank that two elderly individuals requested early withdrawal of fixed deposits.

The Hanyang Branch of the Wuhan Municipal Public Security Bureau received a report from a bank that two elderly individuals requested early withdrawal of fixed deposits. Bank staff suspected they were victims of telecom fraud. Though the two elderly individuals did not know each other, they coincidentally went to the same bank to conduct their transactions.

![2024101401.png](https://cdn.aisecurius.com/images/091/0917d53a-1446-4928-a9dd-d5594a1a8ad0.png)

Upon investigation, Mrs. Xiao and Mr. Yang both received calls from individuals pretending to be their grandsons. Mrs. Xiao was told that her grandson had injured someone in an altercation and was being mediated at a police station, while Mr. Yang’s "grandson" claimed to have been in a traffic accident and needed immediate compensation. Shockingly, both elderly individuals were convinced the voice on the phone sounded exactly like their grandsons.

![2024101402.png](https://cdn.aisecurius.com/images/9cc/9ccb00d4-efe4-44a8-ba64-302ce27304a0.png)

However, after the police confirmed with the families, it was found that neither of their grandsons was involved in any of these emergencies. Police quickly determined this was a new type of fraud in which criminals use AI technology to clone the voices of the elderly's relatives, simulating a real conversation to craft an elaborate scam.


# Analysis of the Scammers' Process


In this type of precise telecom fraud, scammers use AI voice cloning technology to convincingly imitate the voice of a victim's relative, exploiting the elderly's trust and emotional dependence on family members, further enhancing the credibility of the scam. This not only makes the victim more likely to believe the phone call, but also puts them in a state of panic, leading them to make hasty, erroneous decisions.

1. **Obtaining Personal Information**: Scammers first acquire detailed personal information about the elderly's grandsons, including names, contact details, and schooling information, often through illegal data breaches, social media, or other illicit means. With this data in hand, scammers can use AI voice cloning technology to convincingly replicate the voice of the victims' grandsons. By accurately mimicking speech patterns and tones, the scammers make the deception even more believable.

2. **Creating Urgent Scenarios**: In the phone calls, scammers fabricated two different urgent situations — one claimed the grandson was involved in a school altercation, while the other said the grandson was in a traffic accident. Both scenarios played on the elderly's concern for their grandsons' safety, generating a sense of urgency that pressured the victims into reacting quickly. By creating this atmosphere of urgency, scammers manipulate victims into immediately transferring money to resolve the "emergency," emphasizing secrecy to prevent the victims from verifying the story with others.

3. **Exploiting Emotional Dependence**: Elderly individuals often have a strong emotional reliance on their family, especially on their children and grandchildren. When receiving a call from someone imitating a family member, the strong emotional connection can override rational thinking. Scammers exploit this deep-seated trust and care, using it as the point of entry to successfully manipulate victims into making irrational decisions. By deliberately creating a sense of urgency, scammers can cause elderly victims to lose their ability to think critically and act hastily. The elderly, due to their lack of experience in dealing with unexpected crises and their lower adaptability to new technology, are particularly vulnerable to being led by scammers under stress and anxiety.

4. **Key Role of AI Voice Cloning Technology**: AI voice cloning is the most crucial tool in these scams. For the elderly, the cloned voices sound genuine and are difficult to distinguish from the real thing. Most older individuals are not aware of or equipped to handle such sophisticated technological fraud, making them more susceptible to these types of schemes. Additionally, scammers typically isolate the victim, urging them not to inform anyone else, especially family members or bank staff. This isolation tactic minimizes the chances of the victim seeking external help or verifying the legitimacy of the situation, allowing the fraud to succeed more easily.


# How the Elderly Can Prevent AI Phone Scams


In this type of sophisticated telecom fraud, the use of AI-generated realistic audio and video makes it difficult for victims to detect anomalies through traditional methods. Therefore, even after "verifying" the caller's identity via phone or video, the elderly must remain highly vigilant when facing urgent financial requests. When it comes to financial transactions, it's essential to use multiple methods to confirm the identity of the other party and avoid making hasty decisions that could lead to falling into a trap. The **Dingxiang Business Security Intelligence Center** suggests the following effective preventive measures for the elderly:

**1. Redial the Call:**  
When receiving a suspicious phone call or video message, stay calm and avoid immediately responding to any requests. You can use a pretext, like poor signal, to hang up the call. Then, redial using a known, secure contact method to verify whether the request genuinely came from a family member. This prevents you from responding directly to fraudulent information.

**2. Use a Security Keyword:**  
Set up a "security word" or "challenge question" that only you and your close family members know, to verify the caller's identity. If you receive a suspicious call, ask the caller to provide this keyword. If they cannot answer correctly or attempt to avoid the question, this should be a warning signal of a potential scam. Immediately hang up and verify the identity through a secure method.

**3. Consult Family and Friends:**  
If faced with urgent financial requests, first check with other family members or trusted individuals to verify if they are aware of the situation or if they have also received a similar call. This prevents the pressure and confusion caused by handling false information alone. Consulting trusted relatives can help you assess the authenticity of the situation.

**4. Report to the Authorities:**  
Dial the local emergency number, **110**, to report the suspected scam in detail to the police.

These preventive measures not only help individuals defend against AI scams but also remind the elderly to stay calm and vigilant when dealing with sudden financial requests, avoiding falling into the traps set by criminals.


Two elderly people rush to the bank after receiving fraudulent AI-cloned calls

During the National Day holiday, a large number of videos featuring "Lei Jun's AI voice" appeared on short video platforms. In these videos

During the National Day holiday, a large number of videos featuring "Lei Jun's AI voice" appeared on short video platforms. In these videos, "Lei Jun" sharply commented on popular topics such as traffic jams and holidays, and some even contained inappropriate language. Besides the complaints, the "Lei Jun" in the video also claimed to be investing in a certain university.

![2024101103.png](https://cdn.aisecurius.com/images/dab/dabaf1da-1258-4bde-a325-e856947b1946.png)

AI has been widely applied in many fields, from generating images and videos to answering online queries, leading to a wide array of content. When ordinary people can easily use AI to impersonate celebrities and mass-produce audio and video, it inevitably brings problems like information distortion and cognitive confusion to society.

According to *The New York Times*, AI-generated images of Elon Musk have appeared in thousands of fake advertisements, resulting in billions of dollars in fraud cases. An 82-year-old retiree, Tiv Beauchamp, saw a “real interview” with Musk online. In the video, Musk endorsed an aggressive investment opportunity that promised quick returns. The lip movements and voice were highly synchronized, making the video seem very real. The elderly man, believing it to be authentic, contacted the company behind the ad and followed their guidance to set up an account. Over the following weeks, the elderly man spent his entire retirement savings in a series of transactions, ultimately investing over $690,000 (around 4.95 million RMB). Unfortunately, all this money was swindled by scammers, leaving the man devastated.


# These 10 Celebrities Are Frequently Impersonated by Scammers


The advent of AI has completely transformed many industries. Using AI to easily forge the likeness and voice of celebrities is not only used in phishing scams but also to promote fraudulent products, steal personal information, and trick people into downloading malware. Fans often find it difficult to tell if the content is real or fake.

![2024101101.jpg](https://cdn.aisecurius.com/images/2cd/2cdd967b-c455-4e13-b997-bdccee7f564a.jpg)

Social media is an essential platform for marketing and advertising, making it a prime channel for scams. In a report released by the security company McAfee, they listed the 10 most impersonated American celebrities. Criminals exploit their image, voice, and reputation to conduct phishing, cryptocurrency fraud, ticket resale scams, false event promotions, spread misinformation, and promote malware.

- **Scarlett Johansson**: Actor and singer, her name and likeness have been used by scammers for false advertising endorsements.
- **Kylie Jenner**: Reality TV star, her name and likeness have been used by scammers in giveaway scams, fake product promotions, and phishing websites.
- **Taylor Swift**: Singer, her name and likeness have been used by scammers for fake endorsements, ticket scams, giveaways, and spreading false information.
- **Anya Taylor-Joy**: Actor, her name, likeness, and social media accounts have been used by scammers in giveaway scams and spreading false information.
- **Tom Hanks**: Actor, his name and likeness have been used by scammers to promote fake “miracle cures” and fraudulent medical treatments.
- **Sabrina Carpenter**: Singer, her name and likeness have been used by scammers in fake ticket scams and the creation of explicit apps.
- **Sydney Sweeney**: Actor, her name and likeness have been used by scammers in cryptocurrency scams.
- **Blake Lively**: Actor, her name and likeness have been used by scammers in weight-loss gummy scams.
- **Johnny Depp**: Famous actor and director, his name and likeness have been used by scammers for fake endorsements, fraudulent gifts, and cryptocurrency scams.
- **Addison Rae**: Top social media star, her name and likeness have been used by scammers for fake endorsements, fraudulent gifts, and cryptocurrency scams.


# How Ordinary People Can Identify Fake Celebrity Ads


How can ordinary people identify fake celebrity ads? There’s no simple answer, but some essential measures can help maximize the chances of detection.

1. Social media platforms are hotbeds for misinformation and online scams. Be skeptical of shocking or too-good-to-be-true claims, especially when celebrities are involved. Don’t forget to use common sense and logic.
2. Carefully check for unnatural features in the content – strange blinking patterns, mismatched audio, or distorted body parts could be signs of deepfakes.
3. Verify the claims by cross-referencing with trustworthy news outlets and websites. Do not believe content without first performing fact checks.
4. Before making decisions or purchases, consult with friends, family, or colleagues. It’s best to do some deep research and not take the promoter’s word as gospel, even if it’s Elon Musk.
5. Don’t comment, share, or even click on unverified social media posts. Interacting with accounts impersonating celebrities increases your risk of being scammed.
6. Never submit your personal details to unverified individuals, websites, or apps.

### Social Media Platforms’ Defense Against Fake Celebrity Ads

In addressing the problem of fake celebrity ads and information on social media platforms, **Dingxiang Defense Cloud Business Security Intelligence Center** suggests that first, accurately identifying fake accounts is crucial to effectively preventing fraud.

1. **Identify abnormal devices**: Dingxiang's Device Fingerprint technology records and compares device fingerprints to distinguish between legitimate users and potential fraudsters. By uniquely identifying each device, it can detect maliciously controlled devices such as virtual machines, proxy servers, and emulators. The technology analyzes whether devices exhibit abnormal behaviors, such as multi-account logins, frequent IP address changes, or frequent modifications to device attributes, which deviate from user habits. This helps trace and identify fraudsters’ activities.

2. **Identify abnormal account operations**: Remote logins, device switches, phone number changes, and suddenly active dormant accounts require enhanced frequent verification. Additionally, continuous identity verification during sessions is critical. Keeping persistent checks ensures that the user's identity remains consistent throughout use. **Dingxiang’s atbCAPTCHA** quickly and accurately differentiates between human operators and bots, precisely identifying fraudulent activities, monitoring in real-time, and intercepting abnormal behaviors.

3. **Prevent fake videos with face-swapping**: Dingxiang's full-chain, panoramic facial security threat awareness solution intelligently verifies through multiple dimensions, including device environment, facial information, image forgery detection, user behavior, and interaction status. It quickly identifies more than 30 types of malicious attack behaviors, such as injection attacks, live body forgeries, image forgeries, camera hijacking, debugging risks, memory tampering, root/jailbreak, malicious ROMs, and emulators. Once fake videos, false face images, or abnormal interaction behaviors are detected, operations are automatically blocked. The video verification intensity and user-friendliness can be flexibly configured to dynamically adjust the verification for normal users while strengthening it for suspicious ones.

4. **Detect potential fraud threats**: Dingxiang's **Dinsight** helps enterprises conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight’s daily risk control strategy has an average processing speed of under 100 milliseconds and supports the configurable access and accumulation of multi-source data. Based on mature indicators, strategies, and model experience reserves, as well as deep learning technology, it achieves a self-monitoring and self-iterating mechanism for risk control performance. Paired with **Xintell**, a smart model platform, it can automatically optimize security strategies for known risks. Using risk control logs and data to mine potential threats, Xintell allows for one-click configuration of risk control strategies across various scenarios. It standardizes the complex processes of data processing, mining, and machine learning, providing a one-stop modeling service from data processing and feature derivation to final model deployment.

Not Just Lei Jun, These 10 Celebrities Are More Frequently Impersonated by Scammers

Recently, the Chongqing police announced five typical cases of rumors fabricated using AI tools. One of these cases, "

Recently, the Chongqing police announced five typical cases of rumors fabricated using AI tools. One of these cases, "The Hechuan District Public Security Bureau's Handling of Shi for Fabricating a Fake Traffic Violation Ticket with AI Tools," showed that in July 2024, Hechuan District internet user Shi (male, 45) posted a picture of a traffic violation ticket on a short video platform, with a caption stating, "Whistling at a female traffic cop while driving, fined 200 yuan and 11 points deducted." The video attracted widespread attention and discussion among netizens.

![2024101301.png](https://cdn.aisecurius.com/images/4be/4be9a046-25ef-4f7d-acb0-d626b62130c2.png)

The public security agency found that Shi had used AI software to create the fake traffic violation ticket to gain attention, which disrupted public order and had a negative social impact. Shi confessed to the illegal act, and the Hechuan District Public Security Bureau imposed administrative penalties on him according to law.


# How to Identify Internet Rumors

Internet rumors have become an urgent issue that needs addressing in today's online environment. In daily life, individuals need to be vigilant in recognizing and resisting internet rumors.

**Curb Curiosity.**  
When faced with strange or exaggerated information, avoid the temptation to immediately explore further. AI-fabricated rumors are designed to capitalize on trending topics, leveraging the curiosity of netizens to spread false information.

**Verify Unverified Sources.**  
Avoid blindly believing or sharing unverified information. When encountering suspicious content on social media or the internet, first verify it or wait for official announcements before forming conclusions.

**Use Logical Analysis.**  
Pay attention to the logic of online information. If a piece of content has obvious logical flaws or contradictions, it is likely fabricated or manipulated.



# Platforms Strengthen the Detection of Internet Rumors

Rumormongers often use trending news and random online audio/video as material to mass-produce fake news using AI tools. They then distribute it across major online platforms to attract attention or monetize traffic. Some internet platforms, driven by a "traffic-first" mentality, allow fake news, clickbait, and controversial topics to spread unchecked. This creates a favorable environment for AI-generated rumors to thrive, accelerating the formation of a gray industry chain. Content platforms must strengthen management in terms of technology, content, and user accounts.

1. **Strengthen Content Authenticity Identification.**  
Platforms should effectively detect and address content potentially generated by AI to prevent its harmful impact on society. For example, image authenticity detection tools can analyze pixel-level anomalies and imperfections in generated images to determine whether they were created by AI. Additionally, by analyzing user behavior patterns and identity information, platforms can establish security alert mechanisms. Suspicious behavior, such as abnormal logins or frequent information dissemination, can be monitored and restricted. For example, analyzing mouse movement patterns and typing styles can detect abnormal activities, marking deviations from normal user behavior. Additional identity and device verification, along with large models, can quickly sift through massive data sets to identify subtle inconsistencies that are otherwise undetectable by humans, exposing attackers' abnormal operations.

2. **Enhance Multi-Layer Content Verification.**  
To prevent the spread of false information, platforms can introduce multi-layer verification mechanisms for sensitive content. By integrating third-party fact-checking organizations and AI tools, platforms can automatically verify the authenticity of published content. If a risk of falsehood is detected, the platform can label the content as "pending verification" or "suspected misinformation." Additionally, platforms can utilize user reporting systems and feedback from trusted users to help identify rumor content. By incorporating user rating systems, platforms can automatically mark rumored content as low credibility.

3. **Establish Account Reputation Management.**  
Platforms can analyze user behavior patterns and history to identify malicious rumormongers. By monitoring suspicious account activity, such as frequently posting sensitive or false information or widely spreading unverified content, platforms can swiftly identify and restrict rumor-spreading behavior. A user reputation scoring system can be established, where content credibility, reports, and behavioral history adjust the spread of that user's content. High-risk users can be restricted from posting content or receive warning notifications.

Chongqing Police Announce Several Cases of AI-Fabricated Rumors

To protect member data privacy and ensure account security, KrSpace has selected Dingxiang atbCAPTCHA as the tool for registration, login, and identity verification in its member management system.

KrSpace is a leading flexible office service provider. It has creatively launched the KrPass office pass, which offers nationwide access, time-sharing office options, and cost-effectiveness. KrPass is mainly aimed at freelancers, startups, and small teams outside corporate headquarters who require flexible office space. It provides a more efficient office solution. By becoming a KrPass member, individuals can enjoy access to desks, chairs, power and internet, printing equipment, package lockers, property management, cleaning services... as well as free coffee and tea. It meets the daily work needs, client meetings, business dealings, and other diversified business requirements.

![2024101202.png](https://cdn.aisecurius.com/images/557/5574c407-9dbf-44c6-99f2-9a25ca4df6c7.png)

As a crucial platform for KrPass services, KrSpace's member service management system enables nationwide member information synchronization, allowing seamless transitions across KrSpace locations in any city without the need for repeated registration. The self-service cloud printing feature allows members to upload and print documents at any location, while administrators can remotely grant permissions to members through the self-service authorization feature. The self-service invoicing feature meets financial needs nationwide, ensuring regulatory compliance in financial management. Additionally, the system integrates various intelligent access control solutions such as WeChat, app-based QR code scanning, and membership card swiping, providing members with convenient access.

To protect member data privacy and ensure account security, KrSpace has selected Dingxiang atbCAPTCHA as the tool for registration, login, and identity verification in its member management system.

![2024101201.png](https://cdn.aisecurius.com/images/a38/a380cff9-cf5d-4749-a05e-15ea03587d0b.png)

Dingxiang atbCAPTCHA, based on AIGC technology, can prevent AI-based brute-force attacks, automated attacks, phishing threats, and more. It effectively blocks unauthorized access, account hijacking, and malicious operations, ensuring the stability of the system. It integrates 13 authentication methods and multiple defense strategies, compiling 4,380 risk policies and 112 categories of risk intelligence, covering 24 industries and 118 types of risks. Its accuracy rate for risk prevention reaches 99.9%, and it quickly converts risks into actionable intelligence. Furthermore, it allows secure users to pass through without interruption, reducing real-time response time to within 60 seconds. This significantly improves the convenience and efficiency of the account login service, effectively safeguarding user accounts and website information security.

KrSpace operates over 50 communities in 13 cities, including Beijing, Shanghai, Guangzhou, Hangzhou, Nanjing, Suzhou, and Wuhan. The KrPass pass can be used across different communities within the same city and is also valid nationwide, enabling seamless work progress.

Dingxiang atbCAPTCHA Enhances KrSpace Member Account Experience and Security

Recently, the Hangzhou Public Security Bureau disclosed a major online fraud case during a press conference, 

Recently, the Hangzhou Public Security Bureau disclosed a major online fraud case during a press conference, involving the use of AI technology to forge live-action videos, bypassing biometric login authentication on certain platforms. This case highlights how AI technology is being abused by criminals, posing a significant threat to network security defenses.

![2024062706.jpg](https://cdn.aisecurius.com/images/5fd/5fd2b3ee-9bd5-49db-8a42-9813f8d2767e.jpg)

In June this year, Hangzhou cyber police discovered a criminal gang on an overseas platform claiming that they could bypass the biometric authentication of some leading platforms using AI face-swapping technology, illegally obtaining other people's network accounts and information. The gang attracted individuals by promoting services to "help retrieve others' network information," encouraging the use of AI technology for network intrusion activities.

Police investigations revealed that criminals were relying on AI large model technology from overseas sources. By simply uploading a photo, they could generate forged videos through text-based dialogues. These fake videos were used to bypass biometric login procedures on platforms, enabling the criminals to forcibly log into the target user's account and steal personal information.

After obtaining access to others’ accounts, the stolen private data and sensitive information were either sold to fraud groups or used to satisfy some individuals' desire to snoop on others' privacy. Since May this year, the gang had been advertising their "services" on overseas platforms and had successfully stolen large amounts of user data through AI-generated fake identities.

After a thorough investigation, police arrested four key suspects—Hu Mouyun, Hu Mouliang, Zhang Mouguo, and Wu Mouhao—on August 7 in Anhui, Guizhou, and Zhejiang. The gang charged between 2,000 and 5,000 RMB per login attempt, accumulating over 200,000 RMB in illegal profits. The case is still under further investigation.

During the investigation, police found that some platforms' biometric technologies had security vulnerabilities, with relatively simple identification methods, providing opportunities for criminal gangs. In contrast, some platforms had more complex identification systems with stronger defenses against AI-generated videos. This disparity in security technology across platforms highlights the need for further strengthening of login authentication mechanisms to prevent such new types of fraud.

The Hangzhou police advise users to avoid oversharing personal biometric information, such as frontal facial photos and fingerprints, online. Additionally, they recommend regularly monitoring login activities and, if any suspicious logins (such as from unfamiliar devices or IP addresses) are detected, immediately changing account passwords and contacting the platform for assistance.


# User Self-Protection Recommendations


This case not only reveals the risks of AI technology abuse but also reminds both platform operators and users to be vigilant and strengthen information security measures to counter the growing threats posed by emerging technologies.
![2024090402.png](https://cdn.aisecurius.com/images/364/364f34be-178a-41f8-8327-e5249e1cc7aa.png)

Preventing AI fraud requires both technical and behavioral measures to ensure personal information is not easily exposed. Here are specific preventive measures:

1. **Enhance Identity Verification**  
   Enable two-factor authentication (2FA) or multi-factor authentication (MFA), such as SMS verification codes or dynamic passwords, to increase account security. Even if the biometric system is compromised, the additional layers of verification can effectively prevent unauthorized logins.

2. **Limit Sharing of Sensitive Information**  
   Avoid oversharing photos or videos with frontal facial images on social media and other public platforms to prevent AI from using these materials to generate forged information. Only use fingerprint or facial recognition on secure platforms, and choose trusted technology providers.

3. **Regularly Check Account Security**  
   Periodically review the login records provided by platforms to check for any logins from suspicious devices or locations. If any suspicious activity is detected, immediately change your password and notify the platform.

4. **Be Cautious of Suspicious Requests**  
   Even if family or friends "verify" their identity through video or voice, do not easily trust them, especially when financial transactions are involved. Confirm their identity through another channel. Agree with your family or friends on a "security question" that only you know, and use it to verify their identity in case of suspicious requests.

5. **Increase Awareness of AI Technology**  
   Familiarize yourself with common characteristics of AI-generated fake videos or audio. For example, during a video call, ask the other party to perform specific actions, gestures, or expressions to help determine if the video is AI-generated.

By following these preventive measures, individuals can significantly reduce the risk of falling victim to AI fraud. Even as AI technology advances, users can still protect their information by raising security awareness and using reliable security tools.


# Platforms Should Strengthen Prevention


Enterprises can adopt multiple technologies and measures. Additionally, promoting the positive use of AI technology and strictly cracking down on criminal activities is the fundamental solution.

1. **Identify Abnormal Devices**  
   Dingxiang Device Fingerprinting can identify legitimate users and potential fraud by recording and comparing Device Fingerprinting data. It uses unique device identification and recognition technology to detect devices under malicious control, such as virtual machines, proxy servers, simulators, and so on. The system analyzes whether a device is logging into multiple accounts, frequently changing IP addresses, or modifying device attributes, identifying unusual behaviors that don't match user habits. This helps to track and identify fraudulent activities.

2. **Identify Abnormally Operated Accounts**  
   Accounts that log in from different locations, change devices, change phone numbers, or suddenly activate after being dormant need frequent verification. In addition, continuous identity verification during a session is crucial to ensure the user's identity remains consistent throughout the usage period. Dingxiang atbCAPTCHA can quickly and accurately distinguish between human operators and machines, identifying fraudulent activities with precision, and monitoring and intercepting abnormal behaviors in real-time.

3. **Prevent Fake Videos from Face-Swapping**  
   Dingxiang's Full-Chain, Panoramic Face Security Threat Detection Solution verifies information across multiple dimensions, including device environment, facial information, image forgery, user behavior, interaction status, and more. It quickly identifies over 30 types of malicious attacks, such as injection attacks, live body forgery, image forgery, camera hijacking, debugging risks, memory tampering, root/jailbreak, malicious ROM, and simulators. Once fake videos, false face images, or abnormal interaction behaviors are detected, the operation can be automatically blocked. It also flexibly configures the intensity and friendliness of video verification, implementing a dynamic mechanism where normal users are verified with atbCAPTCHA, while suspicious users undergo enhanced verification.

4. **Uncover Potential Fraud Threats**  
   Dingxiang Dinsight helps enterprises with risk assessment, anti-fraud analysis, and real-time monitoring, improving risk control efficiency and accuracy. The average processing speed for Dinsight's daily risk control strategies is within 100 milliseconds. It supports configurable access and accumulation of multiple data sources and can perform self-monitoring and iterative upgrades of risk control mechanisms using deep learning technology. Coupled with the Xintell Intelligent Modeling Platform, known risks can undergo automatic security strategy optimization. Based on risk control logs and data mining of potential risks, it enables one-click configuration to support risk control strategies for different scenarios. Utilizing relational networks and deep learning technology, it standardizes complex processes like data processing, feature derivation, model building, and ultimately deploying models into production, providing an end-to-end modeling service.

Criminal Gang Uses AI to Forge Facial Videos, Successfully Logging into Others’ Accounts

In 2024, during the Ministry of Public Security's "Net Clean 2024" special operation, a case involving the use of AI to create fake obscene images attracted widespread attention.

In 2024, during the Ministry of Public Security's "Net Clean 2024" special operation, a case involving the use of AI to create fake obscene images attracted widespread attention.

![2024100701.png](https://cdn.aisecurius.com/images/a87/a87970a7-31a5-48c7-9006-076f0505444b.png)

The suspect, Liu Moumou, obtained photos of victims by secretly taking pictures and downloading them from social media platforms like "WeChat Moments." He then used AI software to create fake obscene images, incorporating the victim's personal information and spreading the images online, which caused large numbers of netizens to engage in attacks. This seriously impacted the victims' normal work and life. Liu Moumou has now been placed under criminal compulsory measures in accordance with the law.

#   Case Technical Process Analysis

In this case, the suspect, Liu Moumou, carried out the criminal activity through the following technical process:

1. **Stealing Photos**: The suspect first obtained photos through secret photography or by downloading them from social media platforms such as "WeChat Moments." The personal photos shared on these platforms became a "resource library" for the criminal's AI image forgery. Many people frequently share personal life photos on social media, often with weak privacy protection, providing convenient conditions for such criminal acts.

2. **Image Forgery**: Liu Moumou used available AI image generation software to combine the victim's real photos with obscene images, creating fake obscene pictures. By using AI technology to replace faces, adjust colors, and enhance features, highly realistic fake images were produced.

3. **Online Dissemination and Malicious Use**: After creating the fake images, Liu Moumou added the victim's personal information and widely spread the images on social networks, leading to mass attacks and accusations from netizens. This not only caused severe damage to the victim's mental well-being, reputation, and life but also triggered a larger-scale cyberbullying event.


# Preventing the Misuse of Personal Photos


1. **Focus on Personal Privacy Protection**: As social media has become an indispensable part of people's daily lives, avoiding excessive sharing of sensitive information is essential. Social media can become a warehouse of fraud materials for criminals. Avoid sharing personal photos, voices, videos, and other sensitive information on social media. Limit the exposure of private details such as personal accounts, family, or work to reduce the risk of identity theft. If you discover AI-generated fake voices or videos, immediately report them to social media administrators and law enforcement, and take steps to remove them and trace their source.

2. **Strengthen Content Review on Social Platforms**: Social platforms should implement effective identification and timely handling of content potentially involving AI forgery to prevent negative social impacts. For instance, image authenticity detection tools can analyze pixel-level anomalies and detect “flaws” in the image generation process to determine whether the image was AI-generated.

At the same time, social media platforms should establish a security alert system by analyzing user behavior patterns and identity information. They should monitor and restrict suspicious activities, such as abnormal logins and high-frequency messaging. For example, by analyzing mouse movement patterns, typing styles, and other user behavior patterns, platforms can identify abnormal situations, flag suspicious activities that deviate from regular use, and detect attackers' abnormal operations through extra identity and device verification. Large models can quickly filter massive amounts of data and identify subtle inconsistencies that humans may miss.

3. **Dingxiang Device Fingerprinting**: This technology generates a unified and unique device fingerprint for each device. It builds a multi-dimensional identification strategy model based on device, environment, and behavior. This model identifies risky devices, such as virtual machines, proxy servers, and emulators that may be maliciously controlled. It also analyzes whether a device has multiple account logins, frequently changes IP addresses, or alters device properties in ways that show abnormal behavior or deviate from typical user patterns. This helps trace and identify fraudsters’ activities, enabling businesses to manage the same ID across different channels and enhance cross-channel risk identification and control.

4. **Dingxiang Dinsight**: This tool helps businesses conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight’s daily risk control strategies are processed within an average speed of 100 milliseconds. It supports the configuration and accumulation of multi-source data, allowing risk control to self-monitor and iterate using deep learning techniques. Paired with the Xintell intelligent model platform, Dinsight can optimize security strategies for known risks, mine potential risks from risk logs and data, and enable one-click configuration for different scenarios to support risk control strategies. Based on association networks and deep learning technology, the platform standardizes complex processes such as data processing, feature derivation, machine learning, and provides an end-to-end modeling service from data processing to final model deployment.

By leveraging these technologies, social platforms can more effectively screen and filter uploaded content, preventing the spread of fake images.


The Ministry of Public Security Discloses a Case of Using AI to Create Obscene Images

the Gaoxin Branch of Hefei Public Security Bureau released a police briefing, stating that the "Three Sheep" company reported that the online spread of audio and video involving Lu Moumou was false,


On September 26, the Gaoxin Branch of Hefei Public Security Bureau released a police briefing, stating that the "Three Sheep" company reported that the online spread of audio and video involving Lu Moumou was false, severely infringing on their legitimate rights.

Investigations revealed that on the evening of September 22, the suspect Wang Moumou used audio and video materials downloaded from the internet to fabricate a script of Lu Moumou’s alleged drunken remarks. Wang then trained AI tools to generate fake audio mimicking Lu Moumou. The female voice heard in the recordings was also generated using AI tools. Subsequently, Wang used video software to synthesize the audio and video, publishing them online, leading to widespread rumors.

Wang Moumou has been subjected to criminal compulsory measures, and the case is under further investigation.
![2024092701.png](https://cdn.aisecurius.com/images/9f5/9f5c71a9-b418-42a2-a5f3-5f4ec7624567.png)


# Where was Lu Moumou’s Voice Stolen From?


Cloning anyone’s voice using AI technology has become incredibly easy. By extracting voice samples from public sources, even just a few seconds of a sample can be enough to produce highly convincing voice clones. These samples can mimic not only a person’s tone and inflection but also their speech habits and mannerisms, making it extremely difficult for recipients to detect any discrepancies in authenticity.

Dingxiang’s Cloud Business Security Intelligence Center suggests multiple avenues for such information leakage:

1. **Social Media**: Social media has become an integral part of people’s daily lives. Over-sharing sensitive information on these platforms has turned them into a repository for criminals seeking fraudulent materials.

2. **Data Breaches**: Large-scale data breaches may expose personal information, including photos, videos, and audio. Criminals can acquire this leaked data from dark web platforms to carry out fraudulent activities.

3. **Phishing and Malware**: Criminals often use phishing emails or malware to extract personal information, including photos and videos. Once they gain access to a victim's device, they can further obtain additional personal data.

4. **Public Events and Live Gatherings**: At public places like conferences, exhibitions, or social gatherings, criminals may collect photos, voices, and videos of their targets.


# How Individuals Can Prevent Voice Cloning


Facing the challenge of AI voice cloning requires a multi-layered defense. On one hand, it’s crucial to detect and identify forged voices, with platforms needing to strengthen their capabilities to recognize voice cloning. On the other hand, preventing the misuse and spread of cloned voices through user behavior and identity verification is essential to reduce fraud by attackers effectively.

1. **Reduce Sharing of Sensitive Information**: Individuals should avoid sharing personal photos, voice, and video on social media to minimize the risk of identity theft. Limiting the public disclosure of personal accounts, family, and work-related private information reduces the chance of being targeted. If you come across “deepfake” content, immediately report it to social media administrators and law enforcement to have it removed and trace the source.

2. **Platforms Strengthen Voice Cloning Detection**: Social media and communication platforms must enhance their ability to detect voice cloning using advanced AI technology for authenticating voices, quickly identifying, and blocking suspicious content. Platforms should develop and apply AI tools for voice authentication, such as voiceprint recognition, to ensure the security of voice communication.

3. **Identify Abnormal Operations Through Behavior Analysis**: Platforms should analyze user behavior patterns and identity information to establish security warning mechanisms. This involves monitoring and restricting suspicious activities like abnormal logins or high-frequency messaging. By analyzing mouse movement patterns, typing styles, and other behavioral patterns, platforms can flag suspicious activities that deviate from normal usage. Additionally, extra identity and device verification, as well as large models capable of scanning vast amounts of data, can identify subtle inconsistencies that are usually undetectable by humans, helping uncover abnormal operations by attackers.

Dingxiang Device Fingerprinting can generate a unified and unique fingerprint for each device. It builds a multi-dimensional recognition strategy model based on device, environment, and behavior to detect risk devices like virtual machines, proxy servers, and simulators that may be maliciously manipulated. The system tracks and identifies fraudulent activities by analyzing abnormal behaviors, such as frequent IP address changes, multiple account logins, and alterations in device attributes, helping businesses manage risk across channels with a single ID.

Dingxiang atbCAPTCHA is based on AIGC technology and prevents AI brute force cracking, automated attacks, and phishing threats. It effectively blocks unauthorized access, account theft, and malicious operations, ensuring system stability. It integrates 13 verification methods and multiple defense strategies, with 4,380 risk strategies, 112 risk intelligence categories, covering 24 industries, and 118 risk types. Its accuracy rate is as high as 99.9%, quickly converting risk intelligence and reducing response time to under 60 seconds, further enhancing the convenience and efficiency of digital login services.

Dingxiang Dinsight’s real-time risk control engine assists businesses in risk assessment, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk management. Dinsight’s daily risk control strategy processing speed is under 100 milliseconds, supporting the integration and storage of multi-source data. It leverages mature indicators, strategies, model experience, and deep learning to achieve self-monitoring and iteration of risk control performance. Paired with the Xintell intelligent model platform, known risks are automatically optimized, and potential risks are identified through risk control logs and data mining. It offers one-click configuration for various scenarios, supporting risk control strategies. By standardizing the complex processes of data processing, feature derivation, model building, and deployment, Xintell provides an end-to-end modeling service.

4. **AI Voice Tools Should Enforce Anti-Counterfeiting Features**: For instance, incorporating small disturbances, erratic noises, or fixed background rhythms allows listeners to distinguish between real and AI-generated voices. Additionally, hardware that records audio in AI voice synthesis tools should integrate built-in sensors to detect and measure human biological signals when speaking, such as heartbeat, lung movement, vocal cord vibration, as well as the movement of lips, jaws, and tongue. This data can be attached to the audio, providing listeners with verifiable information to distinguish between naturally recorded voices and AI-generated forgeries.

Analysis of the "Three Sheep Recording" Incident: How Criminals Use AI to Synthesize Fake Audio and Video

Recently, Mr. Ma from Guyuan City, Ningxia, reported to local police that he had been a victim of telecom fraud, losing 15,000 yuan to a scammer pretending to be his cousin. 

Recently, Mr. Ma from Guyuan City, Ningxia, reported to local police that he had been a victim of telecom fraud, losing 15,000 yuan to a scammer pretending to be his cousin. This scam not only involved AI face-swapping technology but also exploited the victim's trust in his family members.

The incident began when Mr. Ma was browsing videos on a platform during his free time and suddenly received a message from his "cousin." The person claimed to be working in Africa and, due to the unstable local network, was unable to frequently contact family. The "cousin" then requested to be added as a friend, stating that he needed help with an urgent matter. Trusting his relative, Mr. Ma immediately accepted the friend request without hesitation.

![2024091801.png](https://cdn.aisecurius.com/images/e4e/e4eac092-d3c0-4d03-9e1d-3e621dfd3711.png)

After adding each other, the "cousin" made a video call. In the video, the "cousin" appeared normal, speaking familiarly, with a face that looked similar. He even mentioned some family matters, further dispelling Mr. Ma's doubts. During the chat, the "cousin" claimed that he urgently needed to sign an important contract with his company, but due to the unstable network in Africa, the transfer function could not be used properly. He asked Mr. Ma to temporarily help by covering the contract funds, promising to return the money once the network stabilized.

Based on the "verification" from the video call, Mr. Ma did not doubt the person and believed it was indeed his cousin. Trusting his family member, Mr. Ma quickly agreed to help. Following the "cousin's" instructions, he made three transfers to a designated bank account, totaling 15,000 yuan.

![2024091802.png](https://cdn.aisecurius.com/images/f20/f2024b0e-92bc-49cb-91f3-ae3bb1e6f90d.png)

However, a few days later, when Mr. Ma attempted to contact his cousin again through WeChat to ask for repayment, he sensed something was wrong. His real cousin said that he had never borrowed money from Mr. Ma, nor was he working in Africa. It was at this point that Mr. Ma realized he had been scammed and immediately reported the case to the police.

The police's initial investigation revealed that this was a typical case of modern telecom fraud. The scammer had used advanced AI face-swapping technology to forge the image of the "cousin" in the video to deceive Mr. Ma. Additionally, the scammer utilized voice-cloning technology to imitate the "cousin's" voice and accent, making the scam even more convincing.


# How to protect oneself


In such scams, fraudsters collect their target's voice patterns, speech, or facial information, and then use AI technology to create fake audio, video, or images, producing almost indistinguishable fake content. This increases the victim's trust and makes the scam highly deceptive. They then use excuses such as borrowing money, investments, or emergency rescues to lure the victim's friends or relatives into transferring money or providing sensitive information like bank account passwords.

This case serves as a reminder to the public that when faced with sudden financial requests from family or friends, even if there is a video "verification," one must remain cautious. When money is involved, it's important to confirm the identity through multiple channels. If necessary, direct phone calls or verification through other family members can help avoid falling into a trap.


### 1. Distinguishing during calls: 
During video chats, ask the other person to press their nose or face to observe changes. If it's a real person's nose, pressing it will cause deformation. You can also ask the person to eat or drink to observe facial movements. Alternatively, request odd gestures or facial expressions, such as waving or making a difficult hand gesture, to verify authenticity. During waving, facial data may be disrupted, causing some shaking, flickering, or anomalies.


### 2. Replaying the video:
 When receiving a suspicious video or call, stay calm, hang up under the pretext of a bad signal, and immediately call back through other means to verify the person’s identity, avoiding an immediate response to potential scam content.


### 3. Asking keyword questions: 
Set a "safety word" or "challenge question" that only friends, family, or colleagues would know. Use it to verify the identity of the caller when receiving suspicious communications. If the person cannot provide the correct safety word or avoids the question, hang up immediately and use a known safe contact method to confirm their identity.。


# How Platforms Can Prevent Fraud


It is recommended that enterprises adopt multiple technologies and methods. Moreover, promoting the positive application of AI technology and severely cracking down on criminal activities is the fundamental solution.


### 1. **Identifying suspicious devices**:
 Dingxiang Device Fingerprinting uses the recording and comparison of device fingerprints to distinguish legitimate users from potential fraudulent behavior. It uniquely identifies each device and detects maliciously controlled devices like virtual machines, proxy servers, and emulators. It also analyzes abnormal behaviors such as multi-account logins, frequent IP address changes, and device attribute modifications that deviate from normal user habits, helping to track and identify fraudsters.

### 
2. **Detecting suspicious account activity**: 
For scenarios such as remote logins, device changes, phone number replacements, and sudden activity from dormant accounts, frequent verification needs to be enforced. Additionally, persistent identity verification during sessions is critical to ensure the user's identity remains consistent. Dingxiang's atbCAPTCHA quickly and accurately distinguishes whether the operator is human or machine, precisely identifying fraudulent activities and monitoring and intercepting abnormal behavior in real-time.


### 3. **Preventing fake face-swapped videos**: 
Dingxiang's full-chain panoramic facial security threat detection solution uses multiple dimensions such as device environment, facial information, image forgery detection, user behavior, and interaction state for intelligent verification. It rapidly identifies over 30 types of malicious activities, including injection attacks, live body forgery, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROM, and emulators. Upon detecting fake videos, forged facial images, or abnormal interactions, it can automatically block operations. The system also allows flexible configuration of video verification intensity and user-friendliness, implementing dynamic mechanisms to strengthen verification for abnormal users while maintaining atbCAPTCHA for regular users.


### 4. **Uncovering potential fraud threats**: 
Dingxiang’s Dinsight real-time risk control engine helps enterprises conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight’s average daily processing speed is within 100 milliseconds and supports the configuration and accumulation of multi-source data. It leverages mature indicators, strategies, model expertise, and deep learning technology to achieve self-performance monitoring and self-iteration of risk control mechanisms. Paired with Dingxiang's Xintell intelligent model platform, the solution automatically optimizes security strategies for known risks, extracts potential risks from risk control logs and data, and allows one-click configuration to support risk control strategies in different scenarios. The solution standardizes complex processes such as data processing, feature derivation, and model building through deep learning, providing a one-stop modeling service from data processing to final model deployment.


A man from Ningxia falls into the "African working cousin" scam, losing 15,000 yuan

Dingxiang App Security can also conduct security checks on apps, identifying potential risk vulnerabilities and making targeted fixes

In the wave of digital transformation, Jiahua Technology, as a full-industry chain IoT big data company, is leading industry innovation with its unique model of "developing products with data, providing services with algorithms." By building the "data factory" system, Jiahua Technology is productizing core technologies such as intelligent manufacturing, IoT, blockchain, identity resolution, and artificial intelligence. The company continuously deepens the research of large model technologies, providing strong support for the digital transformation of government and enterprise users.

![2024092002.png](https://cdn.aisecurius.com/images/59a/59aa379b-68e9-4892-9522-3fca70f73321.png)

Jiahua Technology is one of Beijing's first enterprises in the data asset registration pilot program and has received the country's first listed company data asset registration certificate and the first bank data asset pledge loan. This marks a significant breakthrough for Jiahua Technology in the assetization of data, setting a new milestone for the development of the digital economy.

Faced with the increasing complexity of mobile application attacks, Jiahua Technology has shown foresight by deploying Dingxiang App Security across multiple apps. Dingxiang App Security is an adaptive hardening solution based on graph neural network technology. By conducting in-depth analysis and feature extraction of app code, it automatically selects appropriate methods for encryption, obfuscation, and compression, increasing the difficulty of reverse analysis, effectively preventing apps from being cracked, copied, or repackaged by attackers. The use of encryption and obfuscation engines further enhances the security of the apps.

Compliance has become a key factor in app development. For app developers, compliance means strengthening the protection of user data, implementing appropriate security measures and technical means to ensure the security of user privacy. Additionally, developers must comply with relevant laws and regulations, clearly define user permissions and the purpose of information collection, and promptly update privacy policies and user agreements, providing users with transparency and choice.

Dingxiang App Security can also conduct security checks on apps, identifying potential risk vulnerabilities and making targeted fixes. It protects sensitive data, code obfuscation, code integrity, and memory data, preventing security risks from system vulnerabilities at the source, guarding against data leaks, and ensuring app security. It also provides real-time security monitoring of app operations, monitoring terminal devices, operating environments, and user behaviors, helping apps establish a real-time risk monitoring, warning, blocking, and traceability system.

The deployment of Dingxiang App Security not only enhances the security of Jiahua Technology's apps but also provides a solid foundation for its digital construction.

Dingxiang Assists Jiahua Technology in Digital Construction

Recently, employees at a courier company on Youth Road in Kunming City reported a suspicious situation involving a woman attempting to mail gold bars worth over 300,000 yuan

Recently, employees at a courier company on Youth Road in Kunming City reported a suspicious situation involving a woman attempting to mail gold bars worth over 300,000 yuan. Police quickly responded to the scene and successfully prevented a telecommunications fraud case.

Ms. Wang explained that the "best friend" in question was her former college roommate, who had helped her financially during tough times, and they had stayed in touch ever since. The friend later moved abroad, and despite the distance, they continued to communicate regularly through QQ. On the day of the incident, Ms. Wang received a video call request from her friend's QQ account. In the video, her friend asked Ms. Wang to buy 300,000 yuan worth of gold bars and send them to Guangzhou as a wedding gift. Trusting her long-time friend, Ms. Wang agreed to the request and even paid for the gold bars out of her own pocket.
![2024091401.png](https://cdn.aisecurius.com/images/0ea/0ea29d2b-538b-42a1-8b64-d1099dbfb409.png)

Following the police's advice, Ms. Wang immediately asked the "friend" to verify her identity through an international phone call or other means. Shortly after sending this message, Ms. Wang found that her QQ account had been deleted by the other party. It was then that Ms. Wang realized she had narrowly avoided falling into a telecommunications fraud trap that involved AI technology.


#  The Fraudster's Tactics

This is a classic case where fraudsters used AI technology to swap faces and clone voices to impersonate a friend and commit fraud.
![2024080801.png](https://cdn.aisecurius.com/images/d13/d13be02a-8054-4a98-8d45-78065bedb962.png)

In such scams, criminals collect target individuals' speech, facial data, or video, using AI technology to create fake audio, video, or images. They simulate others' voices or appearances to gain trust and then use excuses such as borrowing money, asking for investments, or emergency assistance to trick their victims into transferring funds or providing sensitive information like bank account passwords. Additionally, criminals might use AI to artificially create audio or video of celebrities, experts, or law enforcement personnel, impersonating them to spread false messages and achieve fraudulent goals.

# The Four Phases of the Scam

**Phase 1: Information Gathering.** The first step in the scam involves collecting the victim's personal information. Fraudsters obtain sensitive data such as facial photos, phone numbers, and voice recordings through black market transactions, illegal database leaks, or organized cyberattacks. This information serves as the foundation for their next steps.

**Phase 2: Forging Face Videos and Cloning Voices.** Once they have gathered enough personal information, criminals use software to capture the target's facial features and voice patterns. With AI synthesis technology, they create fake videos and audio clips.

**Phase 3: Identity Impersonation.** The fraudsters begin contacting the victim, mimicking the tone, speech style, and behavior of the person they are impersonating. They communicate through text, voice, or video, gradually lowering the victim's guard and gaining their trust.

**Phase 4: Inducing Victim to Transfer Money or Mail Goods.** Once the victim fully trusts the impersonated identity, the criminals make their move by asking for money transfers, payments for goods, or mailing packages. They often fabricate urgent situations or business needs, pushing the victim to transfer funds to a designated bank account. Under the combined pressure of urgency and trust, the victim often complies without further verification, following the fraudster's instructions.

# Technical Measures to Prevent DEEPFAKE Fraud

To prevent new types of telecommunications fraud involving "DEEPFAKE" and "AI voice cloning," it is essential to verify any suspicious online activity through offline methods. Increasing the duration of communication, using probing techniques, such as asking the other party to perform specific actions to reveal potential flaws, can be helpful. Additionally, companies should adopt multiple technical measures and tools, while promoting the positive application of AI technology and strictly cracking down on criminal behavior, as the fundamental solution.

1. **Identifying face-swapped videos and voices.** During video chats, one can ask the other party to press their nose or face to observe changes in facial features. If the person is real, their nose will deform when pressed. Alternatively, ask the person to eat food or drink water and observe facial movements. Another strategy is to ask them to perform unusual actions or expressions, such as waving or making difficult hand gestures, to detect any anomalies. Waving, for example, may interfere with facial data and cause some shaking, flickering, or other irregularities.

2. **Identifying abnormal devices.** Dingxiang Device Fingerprinting helps distinguish legitimate users from potential fraudsters by tracking and comparing Device Fingerprints. This technology assigns a unique identifier to each device, identifying malicious devices such as virtual machines, proxy servers, and emulators. It also analyzes whether the device has multiple account logins, frequently changes its IP address, or exhibits abnormal behavior inconsistent with typical user habits, aiding in the detection and tracking of fraudulent activities.

3. **Identifying accounts with abnormal operations.** Abnormal behaviors such as logging in from a different location, changing devices, changing phone numbers, or suddenly reactivating dormant accounts require frequent validation. Continuous identity verification during a session is crucial to ensure user identity consistency. Dingxiang's atbCAPTCHA can quickly and accurately distinguish between human and machine operators, precisely identifying fraudulent behavior and monitoring for abnormal activities in real-time.
![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)
4. **Preventing face-swapped fake videos.** Dingxiang's Full-Chain, Full-Perspective Facial Security Threat Detection Solution conducts intelligent verification through multi-dimensional data, including device environment, facial information, image forgery detection, user behavior, and interaction state. This enables rapid detection of injection attacks, fake live bodies, image forgery, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, emulators, and over 30 other malicious attack behaviors. Upon detecting fake videos, fake face images, or abnormal interaction behaviors, it can automatically block operations. Additionally, it allows for flexible video verification intensity and friendliness, dynamically adjusting to provide atbCAPTCHA for regular users and enhanced verification for suspicious users.

5. **Identifying potential fraud threats.** Dingxiang's Dinsight Real-Time Risk Control Engine helps enterprises conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight's daily risk control strategy processing speed averages under 100 milliseconds, supporting configurable access to and storage of multiple data sources. Based on established indicators, strategies, and model expertise, along with deep learning technology, it offers self-monitoring and iterative risk control mechanisms. Complemented by Xintell, an intelligent modeling platform, Dinsight automatically optimizes security strategies for known risks. It leverages risk control logs and data to uncover potential risks, offering one-click configuration for different scenarios and supporting tailored risk control strategies. The platform standardizes the complex processes of data processing, mining, and machine learning, providing end-to-end modeling services from data preparation to model deployment.

Criminals Use AI Technology to Impersonate "College Best Friend," Woman Scammed Out of 5 Gold Bars

enterprises need to combine technical measures, human resources, management mechanisms, and legal regulations to establish a comprehensive defense system that ensures business security and stability.

To tackle the increasingly complex threats posed by AI, enterprises need to combine technical measures, human resources, management mechanisms, and legal regulations to establish a comprehensive defense system that ensures business security and stability.

# Cutting-edge Anti-fraud Technology

Utilizing advanced technological measures to build a solid defense system is the first step in addressing AI risks.


### Advanced Anti-fraud Tools
. Using technologies such as deep learning, natural language processing, and machine learning, enterprises can monitor and analyze transaction data and behavior patterns in real-time to identify anomalies and potential fraudulent activities. For example, behavioral biometrics can differentiate between legitimate users and cybercriminals by monitoring user behavior traits.

![2024090401.jpeg](https://cdn.aisecurius.com/images/b3d/b3d8baf3-0f9f-4c53-9667-17d4ad081f59.jpeg)

Dingxiang Dinsight's real-time risk control engine helps enterprises conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight’s daily risk control strategy processing speed averages under 100 milliseconds, supporting the configurability and accumulation of multi-source data. Based on mature metrics, strategies, and models, along with deep learning technology, it enables self-monitoring and iterative mechanisms for risk control performance. Paired with the Xintell intelligent model platform, it can automatically optimize security strategies for known risks, mining potential risks from risk logs and data, and configuring risk control strategies for different scenarios with a single click. Using associative networks and deep learning, it standardizes complex data processing, mining, and machine learning processes, offering a one-stop modeling service from data processing to model deployment.


### Multi-layered Identity Verification
. Enterprises should adopt multi-factor authentication, biometrics, and Device Fingerprinting technology to enhance the security of user verification. Biometrics combined with device information and behavioral data can provide more precise identity verification, reducing the risk of account hacking or identity theft.

Dingxiang Device Fingerprinting identifies legitimate users and potential fraud by recording and comparing device fingerprints. This technology uniquely identifies each device, detecting virtual machines, proxy servers, emulators, and other maliciously manipulated devices. It analyzes if a device has abnormal behavior, such as multi-account login, frequent IP address changes, or frequent changes in device attributes, helping track and identify fraudulent activities.

Dingxiang atbCAPTCHA quickly and accurately differentiates between humans and machines, precisely identifying fraud behaviors, monitoring, and intercepting anomalies in real-time. atbCAPTCHA, based on AIGC technology, can prevent brute-force attacks, automated attacks, and phishing threats posed by AI. It effectively prevents unauthorized access, account theft, and malicious operations, thereby safeguarding system stability. It integrates 13 verification methods and multiple defense strategies, covering 4,380 risk strategies, 112 types of risk intelligence, across 24 industries and 118 types of risks. Its control accuracy reaches 99.9%, swiftly converting from risk to intelligence. It also supports seamless user authentication without security compromise, with real-time threat handling reduced to under 60 seconds, further improving account login convenience and efficiency, and ensuring the security of user accounts and information.


### Real-time Threat Intelligence Sharing
. Through threat intelligence-sharing platforms across various industries, institutions and enterprises can collaboratively analyze cyber threats and preemptively prevent potential attacks. Meanwhile, continuous AI model optimization and real-time updates enable rapid responses to new risks.


# Professional Anti-fraud Teams


Although technology is at the core of addressing AI risks, it also requires proper human resource allocation and related training. Enterprises need to cultivate and recruit more professionals with skills in AI, cybersecurity, and other areas. Establishing dedicated anti-fraud teams that include security experts, AI engineers, and business risk managers helps build a multidisciplinary collaboration mechanism, ensuring full capability to respond to complex attacks.
![2024090404.jpg](https://cdn.aisecurius.com/images/b8a/b8a8a4aa-cd68-46c9-b4e3-b72caa29021c.jpg)
Additionally, regular internal training is essential to raise overall employee risk awareness, particularly in recognizing AI technologies and cybercrime. Employees, as the first line of defense, play a critical role in identifying suspicious activities, which is key to preventing data breaches and mitigating internal risks.

A note on applying for phone permissions in compliance guidelines for special business scenarios should be added.


# Scientific Security Management Mechanism

The optimization of management mechanisms should be based on the organization’s overall security architecture, ensuring coordination and efficiency between departments. Enterprises should establish comprehensive AI risk assessment systems, regularly evaluating and adjusting potential AI threats. Standardized emergency response processes should be created for possible AI attack scenarios to ensure swift reactions when issues arise.

Business risk management requires close collaboration with departments like legal, compliance, risk management, and operations. A cross-departmental collaboration mechanism ensures that every department can respond quickly to AI risks, forming a cohesive defense.

Enterprises must also strengthen customer data protection measures by establishing strict data access control mechanisms, ensuring that data is encrypted throughout its collection, storage, and use. Furthermore, internal audits and regulatory oversight should be implemented to ensure that data usage complies with relevant legal requirements and privacy standards.


# Legal Safeguards

Effectively addressing AI-related risks requires legal protection. Governments and regulatory bodies play a crucial role in combating AI crimes, and enterprises must ensure their operations comply with regulatory requirements while actively participating in the development of industry standards.

Businesses must adhere to relevant laws and regulations in their respective countries and regions, especially regarding data privacy, AI usage, and crime prevention. Strengthening interaction with regulatory authorities is also essential to ensure that the implementation of new AI technologies does not violate legal regulations.

Given the complexity and rapid evolution of AI applications, companies should actively engage in industry organizations or standardization efforts to promote the standardization of AI technologies in the financial sector. This ensures clear guidelines and regulations during the implementation and application of these technologies.

The cross-border nature of cybercrime requires legal coordination and cooperation between nations. In combating AI-related crimes, enterprises must collaborate closely with law enforcement agencies across different countries to effectively tackle cross-border crimes and money laundering activities.

Building a Comprehensive Defense System to Address AI Risks

. In the past, developing a powerful malware might take months. But with the proliferation of AI tools, attackers can now automatically generate malware in just a few minutes. 

I[n February 2024, the security company Group-IB discovered a malware named "GoldPickaxe," which is the first AI virus software of 2024. ](https://www.aisecurius.com/blog/article/51)This malware tricks users into undergoing facial recognition and submitting identification information. It then uses this information to generate fake videos. These fake facial videos can deceive the biometric systems of banks, allowing attackers to successfully log into the victim's bank accounts. What makes this even more dangerous is that the malware intercepts the bank's SMS verification codes, allowing attackers to perform transfers, make purchases, and even change account passwords.
![2024090404.jpg](https://cdn.aisecurius.com/images/b8a/b8a8a4aa-cd68-46c9-b4e3-b72caa29021c.jpg)
Creating malware is a specialty of AI. In the past, developing a powerful malware might take months. But with the proliferation of AI tools, attackers can now automatically generate malware in just a few minutes. This automated attack tool greatly enhances attack efficiency, and AI's strong understanding of programming languages allows it to quickly discover potential vulnerabilities in software. These automatically generated malware not only increase the speed of attacks but also shorten the reaction time window for defenders. Cybersecurity teams are forced to face more frequent and complex attacks, posing a significant challenge to existing defense mechanisms.

In addition to creating malware, attackers can use AI to conduct in-depth analysis of entire software or system codebases and suggest modifications or replacements. In such cases, attackers can easily implant malicious code, create counterfeit apps, or inject malware into legitimate software. Moreover, AI-generated code often manages to bypass conventional security checks.

AI can also send automated DDoS attacks. Traditional DDoS attacks rely on vast networks of "botnet" devices to initiate attacks, while AI-driven autonomous systems can automatically locate and exploit vulnerable devices, creating even larger botnets. These intelligent systems are not only capable of launching massive DDoS attacks but can also adjust their attack modes in real time based on the target system's response. This means that the attack systems can "learn" the weaknesses in the target's defenses, continuously optimizing their attack strategies to maximize damage.

Moreover, AI can identify potential vulnerabilities in networks and systems. Attackers use AI-based tools to identify vulnerabilities within the IT networks, software, or systems of financial institutions. These tools not only help attackers discover zero-day vulnerabilities but also automatically generate attack scripts targeting specific systems. This technology allows attackers to conduct more precise and targeted attacks, particularly when targeting financial systems, data centers, and other high-value assets.

AI Enables More Rampant Cyber Virus Attacks

As AI becomes more widespread, phishing strategies continue to evolve, and fraud techniques become more complex and covert

Phishing is a classic form of cyber fraud in which attackers often impersonate trusted institutions or individuals to trick victims into revealing sensitive information. Although this type of fraud has been around for decades, it remains one of the most common and effective methods for penetrating companies, stealing data, and conducting financial fraud. Its success lies in the use of social engineering, manipulating human trust, emotions, and impulses to lead the target to make poor decisions.

![2024090202.jpg](https://cdn.aisecurius.com/images/846/846c8a0b-5d43-42f6-82a9-8abce24660b8.jpg)

Phishing attacks are typically carried out via email, fake websites, applications, phone calls, text messages, and social media. These attacks exploit human weaknesses (such as impulsiveness, curiosity, and trust) to lure victims into providing personal information, clicking on malicious links, downloading malware, or transferring funds.

As AI becomes more widespread, phishing strategies continue to evolve, and fraud techniques become more complex and covert.** Fraudsters can use AI to generate realistic and customized phishing content, significantly increasing the deception and success rate of their attacks. According to 2023 data, AI-driven phishing scams have surged by 3,000%, demonstrating the immense threat of this new technology.


# Changes AI Brings to Phishing Attacks


1. **Making Phishing Attacks Easier**: AI dramatically lowers the threshold for phishing attacks, especially in cross-language attacks. Traditional phishing requires attackers to have a certain level of language proficiency, while AIGC can generate high-quality text, voice, and images based on simple prompts and automatically correct spelling and grammatical errors. This means that attackers can generate seamless phishing content in multiple languages, making it easy to target victims across different regions of the world.

2. **Making Phishing Emails More Convincing**: AI can quickly collect, analyze, and integrate publicly available information about an organization or individual, such as social media profiles, news reports, or other public data. Based on this data, AIGC can generate customized “bait files,” making phishing emails more credible and realistic. As AI-generated content is highly intelligent, the typical red flags in phishing emails, such as grammatical errors, spelling issues, or improper formatting, are entirely eliminated, making it difficult for victims to detect.

3. **Diversifying Phishing Methods**: AI can generate not only precise text but also personalized voice and video content. This allows phishing attacks to expand from traditional emails to voice phishing (vishing) and video phishing (deepfakes). Attackers can create simulated phone voices, fake video conferences, or social media interactions, mimicking real voices and images to deceive targets. This diversity in attack methods makes phishing no longer confined to a single text-based channel but more three-dimensional and multi-layered.

4. **Making Phishing Attacks Harder to Detect**: Egress’s 2023 phishing threat trend report shows that in three-quarters of cases, traditional security measures struggle to distinguish AI-generated phishing emails from those written by humans. This is because AI-generated content is highly natural and intelligent, avoiding the typical crude traces found in phishing attacks. Security technologies and systems need further development to meet this challenge, as existing anti-phishing tools are increasingly struggling to detect and defend against such advanced attacks effectively.

5. **Making Phishing Attacks More Targeted**: AI can help attackers create highly personalized attacks by tailoring content based on the target's interests, habits, profession, and social network. By analyzing social networks, public profiles, and other digital footprints, AI can simulate the target's life scenarios, professional environment, and even daily conversation styles to carry out highly precise, targeted attacks. Compared to traditional broad-scope attacks, AI-driven phishing attacks have a significantly higher success rate and destructive power, as their content better aligns with the target's expectations and trust patterns.


#  AI Diversifies Phishing Methods


- **Phishing Email**: Email phishing is the most common form of phishing, often impersonating official emails from well-known financial institutions, government departments, or other legitimate organizations. Attackers typically ask recipients to provide login credentials, bank account information, or click on malicious links to download malware or trojans. The design of the emails often mimics real email templates, even including legitimate companies' logos, contact information, and other visual elements to increase credibility.
![2024090201.png](https://cdn.aisecurius.com/images/1d5/1d59f85c-ab8b-49be-a6ed-6fced40939d4.png)
- **Website Phishing**: Website phishing typically involves creating fake websites almost identical to legitimate ones, with URLs differing by only a single character, making them extremely difficult to distinguish. Users are often tricked into entering personal information, credit card data, or login credentials on these counterfeit websites. Attackers may even use SSL certificates (https) to increase the credibility of their fake websites, deceiving victims into believing they are interacting with a legitimate site.

- **Smishing**: Attackers send fake text messages pretending to be from banks, government agencies, or other trusted organizations, which contain links to malicious websites or induce victims to call scam numbers. Smishing usually employs urgent or tempting content, such as frozen bank accounts or unexpected winnings, to provoke a quick response from users, causing them to overlook careful scrutiny of the link or message.

- **Vishing**: Voice phishing involves attackers directly calling targets or using automated voice messages to scam them. Attackers impersonate bank representatives, law enforcement, or tech support, attempting to obtain sensitive information such as bank account details, passwords, or social security numbers through voice communication. Vishing often emphasizes urgency and authority, trying to pressure victims into disclosing information.

- **Social Media Phishing**: On social media platforms, phishing typically appears as enticing offers, free gifts, or breaking news. Attackers use fake accounts and malicious links to lure users into clicking, leading to the installation of malware or the disclosure of information. Attackers may also use social engineering techniques to impersonate the user's friends or colleagues, further increasing the success rate of the attack.

- **QR Code Phishing**: With the rise of QR codes, malicious QR phishing has become a new attack method. Attackers may distribute malicious QR codes via emails, flyers, or public spaces, luring users to scan and access malicious websites or download malicious applications. Since QR codes do not directly display URLs, users may unknowingly fall into a trap after scanning them.


# How to Prevent AI Phishing Fraud


T[he intelligence bulletin **"DEEPFAKE Threat Research and Security Strategies"** suggests that to prevent and combat AI fraud, ](https://www.aisecurius.com/blog/article/55)it is essential to both effectively detect and identify AI-generated content and to prevent the exploitation and spread of AI fraud. This requires not only technical solutions but also sophisticated psychological strategies and the enhancement of public safety awareness.

1. **Comparison and Identification of Device Information, Geographical Location, and Behavioral Operations**. **Dingxiang Device Fingerprinting** can help identify legitimate users and potential fraudulent behaviors through the recording and comparison of Device Fingerprinting. This technology uniquely identifies and recognizes each device, detecting maliciously controlled devices such as virtual machines, proxy servers, and emulators. It analyzes whether the device shows abnormal behaviors or actions that do not align with the user’s habits, such as logging into multiple accounts, frequently changing IP addresses, or frequently modifying device attributes, thus helping track and identify fraudsters' activities.

2. **Detection of Account Anomalies**. Strengthen frequent authentication in cases of abnormal activities such as account login from different locations, device changes, phone number changes, or sudden activity in dormant accounts. Continuous identity verification during sessions is crucial; persistent checks should be conducted to ensure that the user's identity remains consistent throughout usage. **Dingxiang atbCAPTCHA** can quickly and accurately differentiate between human and machine operators, accurately identifying fraudulent actions and providing real-time monitoring and interception of abnormal behavior.
3. 
![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)

3. **Prevention of Fake Videos and Images**. **Dingxiang's Full-Chain Panorama Facial Security Threat Perception Solution** uses intelligent verification through multidimensional information such as device environment, facial information, image forgery detection, user behavior, and interaction status. It rapidly identifies more than 30 types of malicious attacks, including injection attacks, liveness forgeries, image forgeries, camera hijacking, debugging risks, memory tampering, rooting/jailbreaking, malicious ROMs, emulators, and other system or operational threats. Upon detecting forged videos, fake facial images, or abnormal interaction behavior, it can automatically block operations. The solution also allows flexible configuration of video verification intensity and user-friendliness, implementing a dynamic mechanism to use **atbCAPTCHA** for regular users and strengthen verification for suspicious users.

4. **Uncovering Potential Fraud Threats**. **Dingxiang Dinsight's Real-Time Risk Control Engine** helps companies assess risks, analyze fraud, and conduct real-time monitoring, improving the efficiency and accuracy of risk management. The average processing speed of Dinsight’s daily risk control strategy is within 100 milliseconds, supporting the integration and storage of multi-party data. Based on mature indicators, strategies, and experience in modeling, coupled with deep learning technology, Dinsight implements a self-monitoring and self-iterating mechanism for risk control performance. Together with **Xintell Intelligent Model Platform**, it enables automatic optimization of security strategies for known risks and supports risk control strategies across different scenarios with one-click configuration. Xintell standardizes complex data processing, mining, and machine learning processes, providing an end-to-end modeling service from data processing, feature derivation, model building, to final model deployment.

When Scammers Start Using AI for Phishing

As the digitalization of the financial services industry accelerates, cybercriminals are increasingly leveraging advanced technologies

As the digitalization of the financial services industry accelerates, cybercriminals are increasingly leveraging advanced technologies, particularly AI, to carry out financial fraud. According to BioCatch's *2024 AI, Fraud, and Financial Crime Survey*, the widespread adoption of digital banking during the pandemic has exposed financial systems to greater risks and provided more opportunities for threat actors. AI technology has not only significantly lowered the technical barriers for committing crimes but has also increased the precision and success rate of fraud.
![2024083001.png](https://cdn.aisecurius.com/images/ece/ece8588f-8d57-47c0-8c96-a84df28dcda3.png)


#  AI’s Contribution to Financial Fraud


The report reveals that over 56% of respondents observed an increase in financial crime activity over the past year. Payment fraud, credit card fraud, and identity theft have emerged as the most common types of financial fraud. Among them, synthetic identity fraud has shown especially rapid growth, being recognized as one of the fastest-growing types of financial crime. This type of fraud involves creating fictitious or composite identity information to apply for credit cards, open new accounts, and engage in borrowing or money laundering activities. Alarmingly, traditional fraud detection models have been largely ineffective at identifying the majority of synthetic identities.

![2024083002.png](https://cdn.aisecurius.com/images/10b/10b0bb2a-9ce0-4222-a27e-fb12ec2cbed1.png)

The report notes that nearly 70% of respondents believe that criminals are becoming more skilled at using AI to perpetrate financial fraud. Through AI-powered voice cloning, fraudsters can mimic the voices of legitimate users to bypass identity verification, circumventing traditional security measures. Consequently, 91% of respondents indicated that their organizations are re-evaluating the use of voice verification technologies.

Furthermore, the use of AI in synthetic identity fraud is becoming increasingly widespread. More than 70% of companies detected synthetic identities during the onboarding of new customers last year, with most of these synthetic identities evading detection by traditional fraud tools. This enables criminals to hide within the financial system under false identities and use sophisticated techniques to avoid detection.


# The Importance of Behavioral Biometrics in Fraud Prevention


In response to the increasingly complex methods of fraud, financial institutions are turning to more advanced defense technologies. Behavioral biometric intelligence has emerged as an effective tool for fraud prevention. This technology analyzes users' online behavior, device information, geolocation, and other factors, combined with AI models, to identify and categorize potential fraudulent activity. Behavioral biometrics can not only capture normal user operation patterns but also detect anomalies, such as unusual application flow smoothness or data handling familiarity—often indicative of fraud.

Financial institutions are leveraging behavioral biometric technology to gain insights that traditional fraud prevention tools cannot provide, allowing them to effectively differentiate between legitimate users and cybercriminals. By applying AI technologies, they can proactively identify threats and protect customers while fostering long-lasting trust.

The report also highlights that financial institutions need to share behavioral biometric intelligence more broadly and work closely with regulatory bodies to effectively combat the wave of financial crime driven by AI. In the future, banks and financial institutions will need to continually enhance their technological capabilities to address the challenges posed by AI and safeguard their customers' assets.

BioCatch Report: AI Becomes a Crucial Technology in Financial Fraud

On Android devices, "screen sharing" is implemented through the MediaProjection service. MediaProjection is a technology available for developers on Android 5.0 and above

At the end of July, the National Financial Regulatory Administration issued a risk alert about new types of telecom network fraud, reminding the public to be wary of scams involving "screen sharing" and "AI face-swapping with voice imitation." The notice urged people to enhance their awareness of risk prevention and identification to protect their finances.

![2024073101.PNG](https://cdn.aisecurius.com/images/ed3/ed375e63-abb3-4789-bb18-ecb6cd58c53d.PNG)

The notice pointed out that in "screen sharing" scams, fraudsters use excuses such as "increasing credit card limits," "canceling bad loans under your name," "terminating insurance services," "flight delay refunds," or "offering gifts" to send texts or make phone calls to individuals. They trick consumers into downloading specific chat or online video conference software and instruct them to enable the "screen sharing" feature. This allows the fraudsters to "monitor" the consumer's phone or computer screen in real-time while "guiding" them through actions such as linking bank cards or changing passwords. In doing so, they gain access to personal banking information, passwords, and verification codes, eventually stealing funds from bank accounts.

In June, the Ministry of Public Security released the top ten telecom network fraud cases, one of which also involved "screen sharing" fraud. In May 2024, a woman named Du in Wuxi, Jiangsu, received a video call at home from someone claiming to be a police officer from the criminal investigation team of the Wuxi Public Security Bureau. The person on the video, dressed in uniform, claimed that Du's bank card was involved in a money laundering crime and that she needed to cooperate with an investigation. Du followed the instructions, downloaded meeting software, and enabled screen sharing to allow the "police officer" to verify the funds in her bank account. The fraudster then told Du that she needed to transfer the funds in her account to a "safe account" to prove her innocence. During the process, the "police officer" also instructed Du to take out a 150,000-yuan loan from the bank and transfer it to the "safe account" to show that her financial transactions were normal. It wasn't until her family intervened that Du realized she had been scammed.

In such cases, fraudsters deceive victims into enabling screen sharing on their phones, leading to the exposure of account passwords. Even without directly asking for information, fraudsters can view everything on the victim's phone, including the characters typed when entering passwords and the verification codes received, allowing them to transfer funds from the victim's bank accounts.


# How Attackers Use "Screen Sharing" to Commit Fraud

One of the scenarios where the "screen sharing" feature is applied involves using the internet to project the screen information of one device in real-time to another device. This includes all content displayed on the screen, such as app screens, SMS pop-up notifications, WeChat messages, other app notifications, account and password entry, unlocking procedures, and other real-time operations. Originally designed for user convenience, this feature has become a tool for fraudsters to profit.

According to an analysis by Dingxiang's Cloud Security Intelligence Center, fraudsters use the following four steps to carry out "screen sharing" scams:
![2024071901.png](https://cdn.aisecurius.com/images/94e/94ec54a1-6b3e-4d4a-8d34-81d4fd333514.png)

### Step 1: Obtaining information and gaining trust.

Fraudsters use illegal channels to acquire victim information and impersonate law enforcement, shopping platform customer service, or staff from student loan cancellation services. They accurately match the victim's situation to gain their trust.


### Step 2: Deceiving victims into enabling "screen sharing."

Using certain apps (e.g., QQ, Tencent Meeting, DingTalk) with built-in screen sharing features, fraudsters can monitor the victim's phone operations in real time.


### Step 3: Logging into online banking and monitoring SMS verification codes in real-time.

Fraudsters first instruct the victim to download an app and enable screen sharing. They then use previously obtained bank card details to log into the victim's online banking and initiate a transfer while monitoring incoming SMS verification codes in real-time.


### Step 4: Stealing funds.

Fraudsters initially trick the victim into making a transfer voluntarily. If the victim realizes the scam or hesitates, the fraudster uses the intercepted verification codes to complete the transaction, transferring the funds without the victim's awareness.


#  Dingxiang Device Fingerprinting Can Prevent "Screen Sharing" Risks


On Android devices, "screen sharing" is implemented through the MediaProjection service. MediaProjection is a technology available for developers on Android 5.0 and above, enabling them to capture or record screens. When a user applies to use the MediaProjection service, Android will pop up a permission request. Only after the user clicks "Confirm" can the screen sharing or recording begin.

![2024071903.png](https://cdn.aisecurius.com/images/7f7/7f74b560-d6c2-4c53-b7b5-21ea50d4fabd.png)

Dingxiang Device Fingerprinting can detect the activation of screen sharing through apps like DingTalk and Tencent Meeting in real-time and immediately issue an alert to warn users about potential risks. When MediaProjection is activated, the system automatically creates a virtual display, and Dingxiang Device Fingerprinting detects this additional screen interface, confirming that the device is in a "screen sharing" state. It then promptly reports this status to the risk control center. The risk control center will issue a warning, reminding the user of potential danger and urging them to stay cautious.

Dingxiang Device Fingerprinting collects and analyzes hardware, software, and behavioral data to uniquely identify and recognize each device. This technology can identify maliciously controlled devices such as virtual machines, proxy servers, and emulators. It analyzes whether a device is showing suspicious behavior, such as multiple account logins, frequent IP address changes, or frequent modifications to device attributes that do not align with typical user habits. By tracking and identifying fraudulent activities, it helps prevent fraud. Dingxiang Device Fingerprinting can also be used as an additional factor for identity verification, enhancing the security of user login and transaction processes. By recording and comparing device fingerprints, it can distinguish between legitimate users and potential fraudsters. It supports Android, iOS, H5, WeChat public accounts, and mini-programs, effectively detecting risks such as emulators, rooted/jailbroken devices, system modifications, and injection attacks. It boasts fast detection, efficient risk identification, over 99% stability, and 100% uniqueness.


# How Users Can Prevent "Screen Sharing" Risks


"Screen sharing" is increasingly being used in various forms of fraud. The final step in these scams typically involves asking the victim to use or download video conferencing software that offers "screen sharing" functionality. The goal is to spy on the victim's device to obtain critical information like bank account passwords.

1. **Do not activate the "screen sharing" feature on Android phones without verification.**
2. **Decline any "screen sharing" invitation from strangers.** Do not use the "screen sharing" function with strangers, as they can see all your actions, including when you enter passwords.
3. **Be extra cautious with private information, especially bank account passwords and verification codes.** Never disclose this information to any stranger.

In short, if you receive a suspicious video or phone call, remain calm. You can use the excuse of a bad signal to hang up, then immediately verify the situation through other means to avoid responding directly to possible scam content.

Dingxiang Device Fingerprinting Protects Against Scams Warned by Financial Regulatory Authority

To effectively address these risks, the "3D Vision Developer Community" has adopted Dingxiang atbCAPTCHA as its identity verification tool for user registration and login.


China's 3D vision technology is currently developing rapidly, especially in applications such as artificial intelligence, autonomous driving, industrial robotics, virtual reality (VR), augmented reality (AR), medical imaging, and consumer electronics. With advancements in deep learning, computer vision, and sensing technologies, 3D vision is gradually becoming a core technology in these industries. The number of professionals in the 3D vision field in China has increased significantly, with the number of technical personnel in related fields such as AI, computer vision, and 3D modeling exceeding hundreds of thousands. The specific number is expected to continue to grow as the industry expands. In particular, interdisciplinary talents in deep learning and computer vision have become scarce resources in the market.

![2024082301.png](https://cdn.aisecurius.com/images/b07/b073dd98-ab92-435a-8a18-5987d8e62c5f.png)

The "3D Vision Developer Community" is an open platform launched by Shenzhen Orbbec Technology Co., Ltd., targeting developers from enterprises, universities, and individual tech groups. It aims to accelerate the commercialization of 3D vision technology in smart security, smart home, intelligent retail, commercial display advertising, smart logistics, smart healthcare, and other fields. It also provides systematic learning opportunities for students in related fields, offering a platform for technical exchange with industry experts. This initiative contributes to cultivating a talent pool for China's 3D vision industry.

The "3D Vision Developer Community" includes features from mainstream developer communities such as Odroid, Khadas, Raspberry Pi, NVIDIA, and RealSense. It provides technical documentation, standard SDK downloads, technical forums, industry insights, product and technical training, and other services, offering professional technical support to developers. However, the traditional user registration and login systems in developer communities face security challenges such as weak passwords, phishing, and brute-force attacks. Additionally, fake registrations and account theft not only negatively impact daily operations, user experience, and resource management but also severely damage the community's credibility and reputation, leading to increased legal and security risks. Therefore, community managers need to adopt robust protective measures to prevent and address these issues.

To effectively address these risks, the "3D Vision Developer Community" has adopted Dingxiang atbCAPTCHA as its identity verification tool for user registration and login.

![2024082201.png](https://cdn.aisecurius.com/images/fa3/fa3d5afe-9b4b-41bf-95a7-b974e051d13d.png)

Dingxiang atbCAPTCHA, based on AIGC technology, can prevent threats such as AI-driven brute force attacks, automated attacks, and phishing. It effectively prevents unauthorized access, account theft, and malicious operations, thereby ensuring system stability. It integrates 13 types of verification methods and various protection strategies, gathering 4,380 risk strategies, 112 types of risk intelligence, covering 24 industries and 118 types of risk categories. Its defense accuracy reaches 99.9%, and it can quickly convert risks into actionable intelligence. Additionally, it supports seamless user authentication without user awareness and reduces the real-time response time to within 60 seconds, further enhancing the convenience and efficiency of account login services while effectively safeguarding user accounts and website information security.

Dingxiang atbCAPTCHA Empowers 3D Vision Development Community Security

AISecureus technical experts shared insights on the current prevalent forms of cyber fraud, cutting-edge anti-fraud products, and illustrated the practical effects of anti-fraud measures through

Recently, AWS "Security Day" was held in Hong Kong. Industry experts from AWS, financial institutions, and security companies discussed the role of cybersecurity in the era of AI and its multifaceted impact on the security landscape. AISecureus technical experts shared insights on the current prevalent forms of cyber fraud, cutting-edge anti-fraud products, and illustrated the practical effects of anti-fraud measures through two case studies.

![2024081904.jpg](https://cdn.aisecurius.com/images/b85/b8539dfa-8bd8-426a-8aa5-ee828c438cd8.jpg)

# Four Common Forms of Cyber Fraud


Device fraud, application fraud, account fraud, and marketing fraud are various types of cybercrime in the digital age where fraudsters exploit technology. These activities not only infringe upon users' property and privacy but also distort the fair competition environment in the market. They present significant challenges to enterprises' security defenses and resource allocation, necessitating an increase in security awareness among users and enhanced technical defenses by enterprises to build a safer network environment.

1. **Device Fraud:** Fraudsters use emulators to forge device properties to register a large number of fake accounts. These fake accounts are used for "bonus hunting" and "fake transactions." Through an automated tool known as a "device farm," fraudsters can easily manipulate these accounts in bulk, increasing the concealment and scale of the fraud. Device fraud not only undermines the fairness of corporate promotional activities but also leads to resource wastage. More seriously, this behavior may result in the depletion of resources for real users and even the leakage of sensitive user information.

2. **Application Fraud:** Application fraud involves fraudsters hacking, tampering with, or forging apps or embedding malicious code, and then relaunching these compromised applications. These malicious apps can steal users' accounts, funds, and other sensitive information. Application fraud poses a severe threat to users' property security and may lead to the leakage of users' privacy. As more users rely on mobile applications for daily operations, the harm of this fraud method becomes increasingly significant.

3. **Account Fraud:** Account fraud typically involves cracking account passwords, spoofing logins, and gaining access to victims' accounts to defraud their acquaintances, colleagues, or friends. This type of fraud often utilizes social engineering techniques to deceive the victim into trusting the fraudster, thereby achieving the fraudulent goal. Account fraud can not only lead to direct financial losses for the victim but also result in severe privacy breaches, potentially affecting the victim's personal and professional life.

4. **Marketing Fraud:** Marketing fraud refers to fraudsters using professional tools to register fake accounts, thereby snatching marketing benefits, occupying user resources, and fabricating promotion effects. This not only prevents companies' marketing budgets from generating actual results but also leads to decision-making biases within the company. Such fraudulent behavior not only wastes corporate resources but may also impact the fairness of market competition, leading to the formation of an unhealthy competitive environment.
![2024081602.jpg](https://cdn.aisecurius.com/images/7a1/7a12ca92-8800-47ab-bed1-5573785f4a90.jpg)

# Three Major Anti-Fraud Products


Effectively combating these fraudulent activities requires joint efforts from users and enterprises to raise security awareness and adopt advanced technical measures to prevent and respond to potential risks. Only by doing so can we ensure the safety of users' funds and information in an increasingly complex online environment. AISecureus technical experts provided a detailed introduction to anti-fraud products such as App hardening, CAPTCHA, and Device Fingerprinting.

1. **Identifying Device Risks:** AISecureus Device Fingerprinting works by integrating information from multiple devices, generating a unified and unique Device Fingerprint for each device. A multidimensional identification strategy model based on device, environment, and behavior is established to identify risk devices such as those under the control of virtual machines, proxy servers, or emulators. It analyzes whether devices exhibit abnormal or non-user-like behaviors, such as multiple account logins, frequent IP address changes, or frequent changes in device attributes. This helps trace and identify fraudulent activities, enabling companies to operate under the same ID across all channels and assist in cross-channel risk identification and management.

2. **Ensuring Application Security:** AISecureus's adaptive App hardening, based on Graph Neural Network technology, deeply analyzes and extracts code features, automatically selecting appropriate methods for obfuscation according to the characteristics of different code blocks. This significantly increases the difficulty of reverse engineering while reducing computational performance consumption by 50%. Through its encryption obfuscation engine, the app code is encrypted, obfuscated, and compressed, greatly enhancing the security of the app code and effectively preventing products from being cracked, copied, or repackaged by attackers.

3. **Protecting Account Security and Preventing Malicious Registrations and Logins:** AISecureus atbCAPTCHA, based on AIGC technology, can prevent threats such as AI brute force cracking, automated attacks, and phishing attacks, effectively preventing unauthorized access, account theft, and malicious operations, thereby protecting system stability. It integrates 13 verification methods and various defense strategies, with a collection of 4,380 risk strategies, 112 categories of risk intelligence, covering 24 industries and 118 types of risks. Its defense accuracy reaches 99.9%, and it can quickly transform from risk to intelligence. It also supports seamless user pass-through security, reducing the real-time response and processing time to within 60 seconds, further enhancing the convenience and efficiency of digital login services.


![2024081903.jpg](https://cdn.aisecurius.com/images/f8b/f8b80fb6-4009-4c72-a8db-d860a7885aec.jpg)

# Two Client Case Studies


AISecureus technical experts delved into the practical applications of their technology in the financial sector and the aviation industry, showcasing their outstanding capabilities in addressing security threats.

In one case, a well-known mobile banking application experienced a severe vulnerability attack, where attackers attempted to exploit multiple vulnerabilities across the client, H5 pages, and web end. Notably, through port vulnerabilities, attackers aimed to conduct mass registration, mass login, and mass order grabbing, posing significant security risks to the banking system. To address this challenge, the bank deployed the AISecureus App hardening solution. This solution provided comprehensive encryption protection for the client, stored data, keys, and data transmission through a powerful security SDK. The results showed that after deployment, all attempted intrusions and cracking activities were successfully blocked, ensuring the app's security while maintaining efficient operation.

Another case involved a large airline, whose ticketing system once faced massive malicious bot attacks, leading to the malicious occupation of system resources. Data showed that as much as 90% of search requests were from malicious bots, severely impacting the search experience for regular users. After deploying the AISecureus solution, high-risk search requests quickly dropped from the initial 5% to just 0.008%, medium-risk search requests decreased significantly from 90% to 24.5%, and normal search requests increased from 5% to 75%. These figures not only demonstrated AISecureus's efficiency in countering malicious attacks but also significantly improved user experience and system performance.

This "Security Day" event was the first of its kind held by AWS in Hong Kong in 2024, with the core concept of sharing intelligence. It aimed to provide participants with fundamental knowledge of cloud computing security and share various new technologies to help enterprises enhance security awareness.

AWS Security Day: AISecureus Experts Detail Three Major Anti-Fraud Products

The development of AI voice cloning technology has revolutionized fields such as voice synthesis and personalized services. However


The development of AI voice cloning technology has revolutionized fields such as voice synthesis and personalized services. However, the dual-edged nature of this technology has also led to a rise in fraud activities based on AI voice cloning, with increasingly varied and rampant methods.


# How AI Voice Cloning Fraud Works


1. **Voice Collection:** Fraudsters collect voice samples of their target from social media videos, public speeches, or intercepted phone calls. These samples include voice fragments, tone, and speaking style, among other detailed information.

2. **AI Learning:** Fraudsters use these voice samples to train AI algorithms, enabling the AI to mimic the target’s voice patterns, tone, and characteristics, and even generate new phrases or sentences that the target has never spoken.

3. **Voice Cloning:** Once trained, the AI can generate highly realistic audio that sounds exactly like the target. These cloned voices can then be used to impersonate and deceive victims.

![2024080801.png](https://cdn.aisecurius.com/images/d13/d13be02a-8054-4a98-8d45-78065bedb962.png)


# How Fraudsters Use Cloned Voices


1. **Impersonating Companies or Individuals:** Fraudsters might call or leave voice messages pretending to be trusted entities, such as banks, companies, friends, or family members, to trick victims into sharing personal information or transferring money.

2. **Manipulating Others:** Fraudsters might impersonate the target to manipulate others into taking specific actions, such as transferring funds or disclosing sensitive information.

3. **Conducting Fake Transactions:** Fraudsters may impersonate the target to place orders or conduct transactions over the phone, thereby carrying out fraudulent activities.


#  How to Prevent AI Voice Cloning Scams


As a novel form of fraud, voice cloning poses a serious threat to individual information security. By understanding how these scams operate and implementing effective prevention strategies, we can significantly reduce the risk of falling victim to such scams. The public should enhance their security awareness, remain vigilant, and actively learn prevention techniques to build a safer, healthier online environment. The Dingxiang Defense Cloud Business Security Intelligence Center suggests that individuals can defend themselves against AI voice cloning fraud in the following ways:

1. **Recognize Warning Signs:** When receiving a suspicious video or phone call, maintain a skeptical attitude. Make an excuse about poor signal to hang up and immediately confirm the call by other means to avoid responding directly to potential scam content.

2. **Be Wary of Urgent Requests:** Fraudsters often create a sense of urgency, pressuring victims to act quickly and inducing panic, forcing them to make hasty decisions without thinking. If the caller makes unexpected requests, especially for personal information or financial assistance, remain skeptical. Be cautious of high-pressure tactics and avoid disclosing sensitive information or making immediate payments. Legitimate organizations typically do not demand urgent actions over the phone.

3. **Protect Personal Information:** Never disclose sensitive information over the phone, such as bank account details, ID numbers, or passwords, especially in unsolicited or suspicious calls.

4. **Set Up a "Safe Word":** Establish a "safe word" or "challenge question" known only to close friends, family, and colleagues, to verify the caller’s identity when receiving suspicious calls or messages. If the caller cannot provide the correct "safe word" or evades the question, hang up immediately and confirm through a known, secure contact method.

5. **Reduce Sharing of Sensitive Information:** Avoid sharing personal photos, voice recordings, and videos on social media to reduce the risk of identity forgery. Be cautious about publicly revealing personal, family, and work-related information. If you encounter "deepfake" content, report it to social media administrators and law enforcement immediately, take steps to remove it, and trace its source.

6. **Enhance Cybersecurity Awareness:** Consider using reliable software to protect your devices from malicious activities. Stay informed about the latest voice fraud and cybersecurity threats. Regularly check credible sources for updates on common scam calls and emerging technologies to boost your security awareness.

7. **Report to Authorities Promptly:** If you realize you’ve become a victim of fraud, preserve all related evidence immediately and report it to the police by dialing 110, so they can intervene in a timely manner.

Additionally, platforms should analyze user behavior patterns and identity information to establish security warning mechanisms, monitoring and restricting suspicious activities such as abnormal logins and high-frequency message sending. By analyzing user behavior, such as mouse movement patterns and typing styles, platforms can identify anomalies and flag activities that deviate from normal usage as suspicious. Extra identity and device verification, along with using large models to quickly sift through massive data and detect subtle inconsistencies usually missed by humans, can help identify fraudulent actions by attackers.


 Seven Crucial Measures to Prevent AI Voice Cloning Fraud

During the peak summer travel season, free tickets to popular museums such as the National Museum of China and the Palace Museum have become a resource out of reach for many tourists, 

During the peak summer travel season, free tickets to popular museums such as the National Museum of China and the Palace Museum have become a resource out of reach for many tourists, as scalpers use illegal means to hoard these tickets and resell them at high prices.
![2024081302.png](https://cdn.aisecurius.com/images/a86/a86d502d-cc6d-4b34-b8f7-f63adbd491f1.png)
According to China Central Television, in early August, reporters repeatedly tried to book tickets through the National Museum's mini-program but found that all reservations within the week were fully booked, making it impossible to visit the museum. However, large numbers of tourists were still able to enter these popular attractions each day. Where did their tickets come from? In a random interview with 30 tourists outside the National Museum, it was found that only 2 had successfully booked tickets through official channels, while the remaining 28 obtained their tickets through scalpers at inflated prices.

Reporters then searched for the "National Museum" on several online platforms and found that most discussions were related to ticketing. Scalping was prevalent on many platforms, with some even offering "guided tour" services as a cover for ticket reselling. These services were simply a front; the real purpose was to use technical means to hoard tickets and resell them at high prices.

![2024081303.png](https://cdn.aisecurius.com/images/0a3/0a3c0b96-66cd-414e-8089-7f58aaabff44.png)

Statistics show that not only the National Museum, but also the Palace Museum, Military Museum, and other popular museums, as well as prestigious universities like Peking University and Tsinghua University, have seen this phenomenon.

The Haidian District People's Procuratorate in Beijing handled a scalping case that revealed the underlying tactics. In the case, three individuals—Li, Wang, and Zhang—used ticket-snatching software to secure reservation tickets for well-known tourist attractions nationwide, then resold them at a markup of 80 to 150 yuan. In less than a month during the 2023 summer vacation, they made over 300,000 yuan in illegal profits.

![2024081304.png](https://cdn.aisecurius.com/images/8d2/8d29a6b1-25a7-432f-afda-4943b5bf7624.png)

It was found that the suspects preloaded tourists' names, ID numbers, and other information into the software and used the built-in SMS code platform to receive verification codes in advance. While normal reservations required tourists to enter this information, taking up valuable time, the suspects could receive the codes two minutes earlier, with a validity of up to ten minutes, significantly increasing their chances of securing tickets. Once tickets were released by officials, these scalpers could lock in a large number of tickets within seconds with just one click.


# An In-Depth Investigation into Museum Ticket Scalping


In July 2023, the 19th issue of "Dingxiang Defense Cloud Business Security Intelligence" investigated why tourists were unable to secure museum tickets during the summer vacation. The investigation found that dozens of popular museums nationwide, including the National Museum of China, Nanjing Museum, Hunan Provincial Museum, and Shaanxi History Museum, particularly the trendy ones, had been almost entirely infiltrated by ticket scalpers.

Dingxiang Defense Cloud Business Security Intelligence Center tracked and analyzed that scalpers resell free tickets mainly through three methods: real-name reservation with a surcharge, resale of tickets at higher prices, and selling guided tour packages. The first and third methods were the most common.

**First Method: Reservation with a Surcharge**  
This requires the buyer to provide their real name, ID number, and other personal information. Ticket prices vary depending on the date, with weekdays being relatively cheaper and weekends or holidays more expensive.

**Second Method: Resale of Tickets at Higher Prices**  
Scalpers holding multiple ID cards book tickets under one person’s name and have another person (the tourist) use that ID to enter, repeating the process.

**Third Method: Selling Unauthorized Tour Packages**  
Some travel agency employees and private guides collaborate to register multiple museum accounts using different phone numbers, reserving tickets for various tourists and charging a service fee ranging from 10 to several tens of yuan per ticket. There are also "electronic guide + ticket reservation" or "ticket package" services, with packages including tickets, guide fees, and travel expenses.

![2024081305.png](https://cdn.aisecurius.com/images/970/9705e910-8ad3-4eac-89ee-4fcaa8bd7260.png)

The scalping phenomenon not only affects tourists' visiting experiences but also undermines the fairness of public resources. In these cases, relying solely on the management measures of museums or attractions is far from sufficient. A comprehensive approach involving legal, technical, and market measures is needed to improve regulatory effectiveness and curb the spread of scalping.


# Why Scalpers Can Secure Tickets


Dingxiang Defense Cloud Business Security Intelligence Center found that scalpers often use cheating tools like plug-in software and script programs to snatch tickets.

Ticket grabbing is a race against time. If person A is 1 second faster than person B, person A will be able to purchase the ticket, while person B will miss out. When it comes to placing orders, humans rely on nerve responses, while software operates based on preset processes. Therefore, the running speed of the software far exceeds that of humans, and the success rate of scalping is much higher than that of ordinary tourists.

Although many museums try to regulate scalpers through ticketing rules and technical means—such as blocking frequent reservations based on phone numbers and IP addresses—scalpers continuously update their cheating tools. These scalper-developed ticket-snatching tools can register, log in, and grab tickets in bulk, allowing them to quickly and instantly hoard a large number of tickets. They only need to fill in the identity information, set the quantity, and configure the running time to complete the automatic purchase.


# Preventing Scalpers from Using Software to Snatch Tickets


To prevent scalpers from snatching tickets, simply restricting IPs is of little significance and may accidentally block legitimate tourist reservations. Dingxiang Defense Cloud Business Security Intelligence Center suggests that in addition to restricting IPs and accounts, technical measures should be enhanced to prevent scalpers' cheating software, thereby effectively curbing scalping.


### Ticketing Rules Restrictions


1. **Increase Document Usage Limits**: Each ID number can only be used to make one reservation; each account can reserve tickets for up to 5 people.

2. **Increase Account Usage Limits**: Some museums require that if an account frequently makes reservations within a week and the no-show rate exceeds 50%, the account (phone number) will be restricted from making reservations for 30 days. Some museums also require that accounts that cancel reservations 3 times within 7 days or 5 times within 30 days, or do not enter the venue after booking, will be placed in a "blacklist" for 30 days, prohibiting further reservations, to prevent scalpers from repeatedly using identity information.

3. **Increase Ticket Release Times and Channels**: Releasing tickets irregularly and increasing manual random ticket releases can also reduce the chance of technical ticket snatching to some extent.


### Technical Measures to Prevent Scalping


1. **Detect and Identify Abnormal Ticket Snatching Devices**: Identify whether the client's Device Fingerprinting is legitimate and whether there are risks such as injection, hooks, or emulators. Identify whether the client's Device Fingerprinting is legitimate, quickly identify risks like flashing or modifying devices, rooting, jailbreaking, hijacking injections, etc. Quickly detect multiple activations from the same device, abnormal behavior from the same device associated with an IP, a large number of gatherings from the same IP in a short time, an abnormal proportion of old device models in the same channel, and an abnormal proportion of old operating systems in the same channel.

Dingxiang Device Fingerprinting can identify devices under malicious control such as virtual machines, proxy servers, and emulators, analyze whether the device exhibits abnormal or non-user-friendly behavior like multi-account logins, frequent IP address changes, and frequent changes in device attributes, helping to track and identify fraudulent activities. This allows for early detection of fraudulent behavior, preventing ticket snatching incidents. It can also serve as an additional factor in identity verification, enhancing security during user logins and transactions. By recording and comparing Device Fingerprinting, legitimate users can be distinguished from potential fraudulent activities.

2. **Detect and Block Abnormal Ticket Snatching Accounts**: Use behavior-based strategies to control accounts that switch between multiple accounts on the same device to place orders.

Dingxiang atbCAPTCHA can verify, judge, and block malicious accounts and malicious scraping behaviors in real time at key stages such as registration, login, and query. It is based on AIGC technology and can prevent threats like AI brute force cracking, automated attacks, and phishing attacks, effectively preventing unauthorized access and intercepting web crawlers from stealing data. It integrates 13 verification methods and various defense strategies, supporting seamless security for legitimate users, with real-time response capabilities reduced to within 60 seconds, further improving the convenience and efficiency of the login service experience.

3. **Enhance Risk Identification and Prevention Capabilities**: Establish a dynamic local list management mechanism based on registration data, login data, and activation data, and maintain corresponding black and white list data, including user IDs, phone numbers, devices, etc. After accumulating a certain amount of online data, use risk control data and business sediment data to model registration, login, order, and snatching behaviors, with the model output directly used in risk control strategies.

Dingxiang Dinsight Real-Time Risk Control Engine helps companies with risk assessment, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. The Xintell Intelligent Model Platform paired with Dinsight can automatically optimize security strategies for known risks, identify potential risks based on risk control logs and data mining, and configure risk control strategies for different scenarios with one click.


China Central Television: Scalpers Profiteering from Popular Museum Free Tickets

Recently, the National Financial Regulatory Administration issued a risk warning against new types of telecom network fraud, urging the public to remain vigilant against evolving scams

Recently, the National Financial Regulatory Administration issued a risk warning against new types of telecom network fraud, urging the public to remain vigilant against evolving scams, enhance risk awareness, and develop the ability to identify threats, in order to safeguard their finances. The National Financial Regulatory Administration identified several new types of telecom network fraud, including “screen sharing” scams, “DEEPFAKE” scams, fraudulent online investment schemes, and fake transactions involving online gaming products.

![2024073101.PNG](https://cdn.aisecurius.com/images/ed3/ed375e63-abb3-4789-bb18-ecb6cd58c53d.PNG)

In “DEEPFAKE” fraud, criminals use pretexts such as “online shop customer service,” “marketing promotion,” “recruitment for part-time jobs,” or “dating and relationships” to contact consumers through WeChat, phone calls, and other means to collect voice, speech, or facial information. They then use “face swapping” and “voice imitation” technologies to create fake audio, video, or images that simulate the voice or appearance of others, thereby gaining the victim's trust. Under the guise of borrowing money, investment opportunities, or emergency assistance, they persuade the victim’s relatives or friends to transfer money or provide sensitive information such as bank account passwords, which the criminals then quickly exploit to transfer funds. Additionally, criminals might use deepfake technology to artificially create audio or video content of celebrities, experts, or law enforcement personnel, and spread false information to facilitate fraud.

The police have previously disclosed several cases of telecom network fraud involving DEEPFAKE technology. Recently, in Ordos, Inner Mongolia, a resident named Ms. Li fell victim to a new type of telecom network scam using DEEPFAKE technology. The scammer impersonated Ms. Li’s old classmate “Jia” and established contact with her through WeChat and QQ platforms, using AI technology to fake a video call, thereby gaining her trust. The scammer then requested Ms. Li to transfer money, citing a need for financial assistance, and even sent a fake bank transfer record. Without verifying the recipient, Ms. Li transferred 400,000 yuan to the scammer’s account. Additionally, Mr. Li from Hebei also became a victim of DEEPFAKE telecom fraud, where scammers used a dating app to lure him into a nude chat and download an illegal chat software. They then used AI technology to create a fake pornographic video and extorted 120,000 yuan from him.


# Fraud Using DEEPFAKE Technology


Dingxiang Defense Cloud Business Intelligence Center published a special report titled “DEEPFAKE Threat Research and Security Strategies,” which systematically introduced the industrial chain behind new types of telecom network fraud based on DEEPFAKE technology. New telecom network frauds operate through a complete industrial chain, where the upstream illegally obtains citizens' private information, the midstream customizes scripts to carry out fraud, and the downstream converts fraud into money.

**Stage One: Information Gathering.** The first step of the scam is collecting personal information about the target. Scammers obtain sensitive information such as photos of the victim’s face and phone number through black market transactions, illegal database breaches, or organized cyber-attacks. This information forms the foundation for the scammers’ next move.

**Stage Two: Fake Video Synthesis.** After obtaining personal information, DEEPFAKE technology captures the victim's facial features through software and combines them with AI synthesis techniques to create fake video content.

**Stage Three: Identity Disguise and Communication.** The criminals begin adding the victim’s contact information. They mimic the target’s tone and style, communicating with the victim through text, voice, and video, gradually lowering the victim's guard and gaining their trust.

**Stage Four: Inducing Money Transfers, Loans, and Extortion.** Once the victim fully believes in the scammer's identity, the criminal issues orders for money transfers, loan payments, or goods purchases. They might fabricate urgent situations or business needs to induce the victim to transfer funds to a designated bank account. Under the combined pressure of urgency and trust, the victim often follows the scammer's instructions without further verification.


# Technical Measures to Prevent New Types of Telecom Network Fraud


To prevent “DEEPFAKE” telecom network fraud, it is essential to verify any suspicious activity online by confirming it offline, increasing the duration of communication, and employing probing tactics such as asking the other party to perform specific actions that may reveal inconsistencies. Additionally, companies are advised to adopt multiple technical measures and methods. Furthermore, promoting the positive application of AI technology and cracking down on criminal activities are fundamental solutions.

1. **Identifying DEEPFAKE Fraud Videos.** During video chats, you can ask the other party to press their nose or face to observe changes in facial features. If it's a real person, pressing the nose will cause it to deform. You can also ask them to eat or drink and watch for facial changes. Alternatively, ask them to perform strange actions or facial expressions, such as waving or making a difficult gesture, to distinguish between real and fake. Waving can disrupt facial data, causing some shaking, flickering, or other anomalies.

2. **Comparing and Identifying Device Information, Geographic Location, and Behavioral Operations.** Dingxiang Device Fingerprinting can identify legitimate users and potential fraudulent activities by recording and comparing device fingerprints. This technology assigns a unique identifier to each device, helping detect maliciously controlled devices like virtual machines, proxy servers, and emulators. It also analyzes whether the device exhibits unusual behaviors, such as logging into multiple accounts, frequently changing IP addresses, or altering device properties, which may indicate fraud.

3. **Strengthening Account Identification.** Activities such as remote logins, device changes, phone number changes, and sudden activity from dormant accounts require more frequent verification. Additionally, continuous identity verification during sessions is crucial to maintain consistent user identity throughout usage. Dingxiang atbCAPTCHA can quickly and accurately distinguish between human and machine operators, precisely identify fraudulent activities, and monitor and intercept abnormal behaviors in real-time.

4. **Preventing DEEPFAKE Fake Videos and Images.** Dingxiang's full-chain panoramic facial security threat perception solution performs intelligent verification using multidimensional information, including device environment, facial information, image forgery detection, user behavior, and interaction status. This solution can quickly identify and block more than 30 types of malicious activities such as injection attacks, live body forgeries, image forgery, camera hijacking, debugging risks, memory tampering, Root/jailbreak, malicious ROM, and emulator operations. Upon detecting forged videos, fake facial images, or abnormal interactions, it can automatically block the operations. Additionally, the solution allows for flexible configuration of video verification strength and user-friendliness, implementing a dynamic mechanism that applies atbCAPTCHA to normal users while strengthening verification for abnormal users.

5. **Uncovering Potential Fraud Threats.** Dingxiang Dinsight real-time risk control engine helps companies with risk assessment, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight’s average processing speed for daily risk control strategies is within 100 milliseconds and supports the configurable access and accumulation of multi-source data. It leverages mature indicators, strategies, model experience, and deep learning technology to achieve self-performance monitoring and iterative risk control. Paired with the Xintell intelligent model platform, Dinsight can automatically optimize security strategies for known risks, mine potential risks from risk control logs and data, and configure risk control strategies for different scenarios with one click. The platform standardizes complex processes such as data processing, mining, and machine learning, providing an end-to-end modeling service from data processing and feature derivation to model construction and deployment.


Financial Regulatory Authority: Alert on “DEEPFAKE” Fraud

As a trillion-dollar industry, used car trading platforms offer consumers diversity and convenience, with a wide selection of used cars and streamlined transaction services,


In recent years, the rapid growth of China's automotive market has fueled the expansion of the used car sector, driving new consumer momentum at an unprecedented pace. According to data from the China Automobile Dealers Association, in the first half of 2024, cumulative used car transactions reached 9.3829 million units, a year-on-year increase of 7.01%, with a total transaction value of 625.212 billion yuan.

![2024080901.jpg](https://cdn.aisecurius.com/images/ac0/ac04dc16-b1ee-4270-a5df-6cf875067a0b.jpg)

As a trillion-dollar industry, used car trading platforms offer consumers diversity and convenience, with a wide selection of used cars and streamlined transaction services, significantly simplifying the car-buying process and improving efficiency.

Che300 is an independent third-party platform for automotive transaction and financial big data services, providing one-stop services for key data in the automotive transaction and financial industry chain. It offers online information query services such as vehicle price evaluation, vehicle maintenance records, model identification, and vehicle value retention analysis for car buyers and sellers.

Che300 enhances the standardization of vehicle model information, vehicle condition information, and vehicle pricing, providing foundational database services for auto dealers and trading platforms. The platform's internet-based comprehensive data search and analysis capabilities reduce potential profit risks in used car transactions. Additionally, it offers vehicle valuation through extensive historical transaction data and provides market analysis for target models. Che300 also supports banks, insurance companies, leasing firms, auto manufacturers, and automotive finance companies with fraud detection, thorough screening, and continuous monitoring functions.
![2024080501.png](https://cdn.aisecurius.com/images/648/648cc32e-f5d5-471f-a89d-5405a78f4b20.png)
Data is a crucial digital asset. To ensure user account security and prevent malicious bot attacks, Che300 has chosen Dingxiang atbCAPTCHA. This tool effectively blocks malicious bots from stealing data and can perform real-time verification, assessment, and interception of malicious accounts and data scraping behaviors during critical stages such as registration, login, and queries.

Dingxiang atbCAPTCHA, based on AIGC technology, prevents AI-driven brute force attacks, automated attacks, and phishing threats, effectively protecting against unauthorized access and intercepting web crawlers. It integrates 13 verification methods and multiple defense strategies, supporting seamless user verification while reducing real-time response capability to within 60 seconds, thereby enhancing the convenience and efficiency of the login experience.

Moreover, Dingxiang atbCAPTCHA can effectively prevent these automated tools from malicious program or personnel mass registrations and logins. Malicious programs or individuals often use brute force attempts to log into website accounts, but Dingxiang atbCAPTCHA significantly increases the difficulty of such attacks, preventing unauthorized logins and safeguarding users' privacy and financial security.

Dingxiang atbCAPTCHA Enhances Data Security in the Used Car Market

accounting for 25.90% and 18.50%, respectively. Additionally, "impersonating leaders," "guess who I am," and cryptocurrency fraud also represented a significant proportion.

Li's QQ account was added to a work group by a scammer. Seeing that the names in the group were those of his company's staff, he did not verify further. Days later, Li received a group message: the scammer, pretending to be the "general manager," claimed that a payment to a contractor was needed and asked Li to check the company's account balance. After verifying the funds, the "general manager" asked Li to transfer all the money to a specified account, urging him to hurry due to the urgency. Fearing to offend the "leader," Li transferred 400,000 yuan from the company account. It was only after the real general manager received a bank notification that Li realized he had been scammed.

![2024072401.png](https://cdn.aisecurius.com/images/040/040acdda-1de2-4c3c-a7aa-10f3413c0a4d.png)

Recently, the Ministry of Public Security announced the top ten types of telecom network fraud, with "impersonating leaders and acquaintances" standing out. Scammers use photos and names of victims' leaders or acquaintances to create fake social media accounts, add victims as friends, or include them in chat groups. They then impersonate leaders or acquaintances to gain trust by showing concern and giving commands in a familiar tone. Claiming to have transferred funds to the victim’s account and asking for a re-transfer to another account, they also send fake screenshots of successful transfers. Trusting the "leader" or "acquaintance," victims often do not verify the identity and believe the funds were transferred to their account. The scammers then push for a quick re-transfer, thus defrauding the victims. This type of fraud exploits the victim’s trust in familiar individuals, neglecting identity verification.


#  Criminals Using AI to Impersonate Leaders


Analysis by Dingxiang Defense Cloud Business Security Intelligence Center shows that from 2015 to 2023, impersonating online shopping customer service and rebate fraud were the most common types of telecom network fraud, accounting for 25.90% and 18.50%, respectively. Additionally, "impersonating leaders," "guess who I am," and cryptocurrency fraud also represented a significant proportion.

![2024072301.png](https://cdn.aisecurius.com/images/212/2122dc68-d940-4be8-b5a0-7f64bbb97dc3.png)

With the advancement of AI technology, high-tech fraud methods like voice cloning have become rampant. "Impersonating leaders and acquaintances" frauds have become more deceptive. Criminals use voice cloning technology and AI algorithms to mimic specific individuals' voices, creating realistic audio that sounds like genuine conversations. This technology typically employs text-to-speech conversion, turning written text into voice output. Scammers combine synthesized voices and fake video calls to further enhance the credibility of their scams. These new telecom fraud methods are more concealed and challenging to prevent.

 
# Four Steps of AI Voice Cloning Fraud


Dingxiang Defense Cloud Business Security Intelligence Center's analysis reveals the four stages scammers use to conduct voice cloning fraud. Each stage is a meticulously designed trap, and if victims cannot effectively discern, they may fall step by step into the scammers' trap.

#### Stage One: Information Gathering

The first step in fraud is gathering personal information about the target. Scammers obtain sensitive information like facial photos, phone numbers, and bank card numbers through black market transactions, illegal database leaks, or organized cyber attacks. This information forms the basis for the scammers' next actions.

#### Stage Two: Voice Synthesis

With personal information in hand, scammers use advanced voice synthesis technology and deep learning algorithms to create realistic voice samples. These samples not only mimic the target's tone and inflection but also replicate their language habits and expressions, making the voice sound indistinguishable from the real person and greatly enhancing the deception.

#### Stage Three: Identity Masquerade and Communication

Once the voice samples are ready, criminals start contacting the victim. They impersonate the target’s tone and style, communicating through text, voice, or video. During this stage, scammers strive to emulate the target’s everyday communication style to lower the victim’s guard and build trust.

#### Stage Four: Inducing Money Transfer

When the victim is convinced of the scammer's identity, the criminals begin issuing commands to transfer money or make payments. They may fabricate urgent situations or business needs, prompting the victim to transfer funds to designated bank accounts. Driven by urgency and trust, victims often comply without further verification, following the scammers’ instructions to transfer the funds.


# How to Prevent "Impersonating Leaders" Scams?


#### Personal Precautions

When you receive a friend request on social media from someone claiming to be your leader, always verify their identity first. Avoid trusting blindly and handle friend requests cautiously. For any requests involving loans or money transfers, especially those that seem urgent or unusual, be vigilant and confirm with the leader in person through reliable channels. Only proceed with financial transactions after thoroughly verifying the situation.

Additionally, do not disclose your personal information to unfamiliar individuals online to enhance your personal prevention awareness. If you find yourself a victim of fraud, immediately preserve all related evidence and call 110 to report to the police so they can intervene promptly.

#### Technical Precautions

Platforms should analyze user behavior patterns and identity information to establish a security alert mechanism, monitoring and restricting suspicious activities such as abnormal logins and high-frequency messaging. By analyzing mouse movement patterns, typing styles, and other user behavior patterns, anomalies can be identified, and activities deviating from normal usage can be flagged as suspicious. Additional identity and device verification measures should be implemented, and large models can quickly screen vast amounts of data to detect subtle inconsistencies usually undetectable by humans, identifying attackers' abnormal operations.

**Dingxiang Device Fingerprinting** can generate a unified and unique Device Fingerprint for each device, identifying risk devices such as virtual machines, proxy servers, and emulators being maliciously controlled. It analyzes whether the device has multiple account logins, frequently changes IP addresses, or frequently changes device properties, identifying behaviors that are abnormal or inconsistent with user habits, tracking and identifying fraudsters' activities.

**Dingxiang atbCAPTCHA** based on AIGC technology can prevent brute force attacks, automated attacks, and phishing attacks by AI, effectively preventing unauthorized access, account hijacking, and malicious operations, thereby protecting system stability.

**Dingxiang Dinsight Real-Time Risk Control Engine** helps enterprises conduct risk assessments, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Paired with Dinsight, the **Xintell Smart Model Platform** can automatically optimize security strategies for known risks, mine potential risks based on risk control logs and data, and configure different scene-supported risk control strategies with one click.


 When Criminals Use AI to Impersonate Leaders for Fraud

Through advanced speech synthesis and deep learning algorithms, scammers generate extremely realistic voice samples that are almost indistinguishable from the real thing

An office worker received an urgent call during a busy workday. The caller claimed to be his boss, informing him that he forgot to transfer funds to a new partner before leaving and needed to handle it immediately. The voice sounded very realistic, and the urgent content made him accept the transfer task without hesitation. After the call ended, he completed the transfer according to the provided wire transfer information. However, it was a scam call based on voice cloning.

![2024071003.png](https://cdn.aisecurius.com/images/249/249445af-4757-47af-b07a-957a6739bc28.png)

The key behind this phone fraud is AI technology. Through advanced speech synthesis and deep learning algorithms, scammers generate extremely realistic voice samples that are almost indistinguishable from the real thing. These voice samples not only mimic a person's tone and intonation but also replicate their language habits and expression styles, making it difficult for the recipient to doubt their authenticity.

The application of voice cloning in phone scams shows that traditional identity verification methods are no longer secure and reliable. Faced with this challenge, technical experts and enterprises need to take multi-layered defense measures: on the one hand, they need to effectively detect and intercept forged voices; platforms need to strengthen the identification of voice cloning; on the other hand, they must prevent the use and spread of voice cloning and verify user behavior and identity; thereby effectively reducing fraud actions by attackers.



# Several Technologies for Identifying Voice Cloning


Using advanced AI technology to authenticate the authenticity of voices, timely detecting, and intercepting suspicious content. Developing and applying AI tools for voice anti-counterfeiting, such as using voiceprint recognition technology to confirm the authenticity of the voice, providing security for voice communication.

### 1. Resemble's Detect-2B

Traditionally, most AI-generated audio clips sound "too clean," lacking the natural noise of real recordings. AI-based models can make judgments by focusing on these subtle differences.

Resemble recently launched the Detect-2B model, an AI solution specifically designed to detect cloned audio. The model architecture is based on Mamba-SSM or state space models, opting not to rely on static data or repetitive patterns but instead using random or probabilistic models. It responds better to different variables and dynamics, maintaining efficiency and accuracy even in poor recording conditions. It effectively handles audio signal diversity, capturing dynamic changes in audio clips, and thus effectively detecting voice cloning.

### 2. Meta's AudioSeal

Meta's Fundamental AI Research (FAIR) team introduced AudioSeal, an audio watermarking technology designed specifically for detecting AI-generated speech. AudioSeal's core technologies include advanced audio feature extraction and comparison algorithms, as well as efficient audio processing and watermark embedding techniques. Through these technical means, AI models can accurately identify and mark AI-generated audio segments, making the detection process faster and more efficient. Reportedly, the new localized detection method increased detection speed by 485 times, greatly enhancing the capability to handle large-scale audio data.

Compared to traditional methods, AudioSeal can accurately locate AI-generated speech segments in longer audio clips, effectively preventing cloning or tampering.

### 3. McAfee's Project Mockingbird

McAfee developed the AI model Project Mockingbird, specifically for detecting and identifying AI-generated audio content. Project Mockingbird uses advanced neural network architecture and large-scale data training. The model can accurately analyze audio features, distinguishing AI-generated audio from real recordings. It has rapid response capabilities, allowing real-time detection and identification during audio streaming, effectively addressing immediate audio fraud risks.

### 4. Multi-Factor Authentication Technology

Based on user behavior and identity to identify attackers. Platforms should analyze user behavior patterns and identity information to establish security warning mechanisms, monitor and restrict suspicious behaviors such as abnormal logins and high-frequency message sending. They should also analyze user behavior patterns like mouse movement and typing style to identify abnormalities, flagging suspicious activities deviating from normal usage. Additionally, extra identity and device verification can be employed. Large models can quickly sift through massive data and identify subtle inconsistencies usually undetectable by humans, discovering abnormal operations by attackers.

**Dingxiang Device Fingerprinting** can generate a unified and unique device fingerprint for each device, identifying risk devices like virtual machines, proxy servers, and emulators manipulated maliciously. It analyzes whether the device has multiple account logins, frequently changes IP addresses, or frequently changes device attributes to detect abnormal or user-unusual behaviors, tracking and identifying fraudsters' activities.

**Dingxiang atbCAPTCHA**, based on AIGC technology, can prevent brute force cracking, automated attacks, and phishing attacks by AI, effectively preventing unauthorized access, account hijacking, and malicious operations, thus protecting system stability.

**Dingxiang Dinsight Real-Time Risk Control Engine** helps enterprises with risk assessment, anti-fraud analysis, and real-time monitoring, improving risk control efficiency and accuracy. Paired with the **Xintell Smart Model Platform**, it can automatically optimize security strategies for known risks, mine potential risks based on risk control logs and data, and configure risk control strategies for different scenarios with one click.


Voice cloning technology presents both opportunities and threats. The development of effective detection and analysis technologies, as well as comprehensive multi-factor authentication mechanisms, are crucial in mitigating the risks associated with AI-generated voice fraud.

Recommended: Mainstream Voice Cloning Detection and Analysis Technologies

On April 23, 2024, the Beijing Internet Court ruled that the defendant tech company and software company should apologize to the plaintiff, 

With the widespread adoption of AI technology, extracting sufficient voice samples of a person can enable the "cloning" of their voice to create AI voice products. This technological advancement also poses challenges to the protection of voice rights.

The first nationwide AI voice infringement case, judged by the Beijing Internet Court, has sparked widespread attention to the boundaries of AI technology application. The case involves Ms. Yin, whose voice was cloned and commercialized without authorization.


# A Legal Case of Voice Cloning


Ms. Yin is a professional voice actress who discovered that her voice was being used for AI voiceover on a short video platform without her authorization. Through investigation, Ms. Yin found that her voice was recorded during cooperation with a cultural media company and then provided to a software company without her consent, ultimately purchased and commercialized by a tech company.

![2024071503.jpg](https://cdn.aisecurius.com/images/2eb/2ebab65b-476f-436b-8bba-766ec3a72308.jpg)

The tech company used AI technology to clone Ms. Yin's voice for a text-to-speech product called "Mo Xiaoxuan," which they sold on their software platform. Ms. Yin believed this action infringed on her voice rights and sued the relevant companies, demanding they cease the infringement, apologize, and compensate her for economic and emotional damages totaling 600,000 yuan. This case not only highlights the conflict between technology and law but also raises a deep reflection on the ethics of AI technology application. The development of voice cloning technology has brought convenience to areas such as speech synthesis and personalized services, but it must be cautiously applied within a legal framework to respect and protect individuals' personality rights.

A key focus of the case is the issue of authorization. The defendants argued that the AI-generated voice lacked identifiability and should not be protected under voice rights. The plaintiff contended that even if the defendants had copyright authorization, it did not equate to authorization of personality rights. The court ruled that if the public can associate the voice with a specific individual based on its timbre, tone, and speaking style, the voice is identifiable and should be protected. The court ultimately determined that the cultural media company and software company used the plaintiff's voice without permission, constituting infringement.

According to Article 1023 of the Civil Code of the People's Republic of China, the protection of a natural person's voice rights follows the relevant provisions of portrait rights. Voice rights are part of personality rights, with the prerequisite for protection being "identifiability."

![2024071504.jpg](https://cdn.aisecurius.com/images/9bf/9bf194cd-2c8e-4f35-bdb2-c270ac6f2bd3.jpg)

Beijing Internet Court's Vice President Zhao Ruigang noted that the iconic nature of a voice is very clear and can be associated with a person, thus voice rights are a new type of iconic personality right added to the Civil Code. Zhu Wei, Deputy Director of the Communication Law Research Center at China University of Political Science and Law, pointed out that voice rights are protected by referencing portrait rights, with the right to recognize a specific person through their voice belonging to specific personality rights. Experts believe that obtaining the legitimate authorization of the rights holder is key to determining voice rights infringement.

On April 23, 2024, the Beijing Internet Court ruled that the defendant tech company and software company should apologize to the plaintiff, and the cultural media company and software company should compensate the plaintiff 250,000 yuan in economic damages. The court's judgment reflects the importance of protecting voice rights and sets legal requirements for the commercial application of voice cloning technology.


#  Where Is an Individual's Voice Stolen?


The development of voice cloning technology has brought revolutionary changes to fields such as speech synthesis and personalized services. However, the dual-edged nature of technology has also led to an increase in scams based on voice cloning. Criminals use voice cloning for fraud in various ways, becoming increasingly rampant.

![2024071003.png](https://cdn.aisecurius.com/images/249/249445af-4757-47af-b07a-957a6739bc28.png)

Cloning anyone's voice based on AI technology is very easy. It only requires extracting voice samples from public channels, sometimes needing just a few seconds or even seconds of samples, to create highly realistic voice clones. These voice samples can not only mimic a person's tone and intonation but also replicate their language habits and expression styles, making it difficult for recipients to doubt their authenticity.

Before implementing new types of telecom network fraud using voice cloning, fraudsters first collect target information, including images, contact details, home addresses, work information, and lifestyle details. The Dingxiang Defense Cloud Business Security Intelligence Center believes that there are various channels through which this information can be leaked:

1. **Social Media**: As social media has become an indispensable part of people's daily lives, avoiding the oversharing of sensitive information on social media has become a repository for fraudsters to gather scam material.

2. **Network Data Breaches**: Large-scale network data breaches can lead to the leakage of personal information, including photos, videos, and voices. Fraudsters can acquire these leaked data on dark web platforms to carry out fraudulent activities.

3. **Phishing and Malware**: Fraudsters send phishing emails or use malware to obtain victims' personal information, including photos and videos. Once they successfully gain access to the victim's device, they can further acquire more personal data. For instance, on February 15, 2024, the foreign security company Group-IB announced the discovery of malicious software named "GoldPickaxe." This malware's iOS version deceived users into face recognition and submitting ID documents, then conducted "AI face-swapping" fraud based on the user's facial information.

4. **Public Events and On-Site Activities**: At public places or specific events such as conferences, exhibitions, and social activities, fraudsters might collect victims' photos, voices, and videos.

In the face of voice cloning challenges, technical experts and companies need to adopt multi-layered defense measures: on one hand, they need to effectively detect and identify fake voices; platforms need to strengthen the identification of voice cloning; on the other hand, prevent the use and spread of voice cloning, verifying the behavior and identity of users; thereby effectively reducing fraudulent actions by attackers. Particularly, it is important to avoid sharing personal photos, voices, and videos on social media, and reduce the exposure of personal accounts, family, work, and other private information to lower the risk of identity forgery. Once discovering fake content of voice cloning, it should be reported immediately to social media administrators and law enforcement to take measures to delete and trace the source.

What Are the Culprits Behind Personal Voice Cloning?

To address the new security challenges brought by AI, Hanxiaocircle App decided to adopt Dingxiang iOS reinforcement.



With the widespread application of AI technology, the security threats it brings have increasingly attracted people's attention. Especially in the mobile internet environment, users' personal information and data security are facing unprecedented challenges.[ In February 2024, a malicious software named "GoldPickaxe" appeared. The iOS version of this software deceived users into face recognition and submitting ID documents, using deepfake technology to create fake face videos, thus defrauding users' bank account information, posing a significant security risk to users.](https://www.aisecurius.com/blog/article/51)

This AI-based new threat not only poses a threat to users' personal information and data security but also could seriously impact users' safety and privacy rights. The development of AI makes false information and fraud easier. Attackers may use this technology to create fake videos or images, damaging the reputation of individuals or organizations, and even causing social chaos. Additionally, social engineering attacks are becoming more rampant, with attackers using AI to generate fake accounts or information, conducting phishing and online scams, leading to privacy leaks and financial losses. To address the new security challenges brought by AI, Hanxiaocircle App decided to adopt Dingxiang iOS reinforcement.

![2024072603.png](https://cdn.aisecurius.com/images/e56/e5666740-9b91-47a5-9e8d-5759ace5d2b4.png)

Hanxiaocircle App is an application that provides a comprehensive collection of content related to Korean dramas and entertainment. It has won the favor of many users with its rich content and smooth user experience. The app offers a vast amount of Korean drama and entertainment content, allowing users to get all the information they are interested in at once. With a large collection of drama video resources, users can enjoy new and popular dramas, with highly-rated shows all available. The smooth interactive experience allows users to quickly play high-definition videos with easy swiping, enjoying the ultimate viewing experience. However, with the increase in user numbers and the diversification of content, the app also faces many security challenges. Particularly with increasingly complex attack methods targeting mobile applications, security protection becomes extremely important.

Dingxiang iOS reinforcement is an adaptive reinforcement solution based on graph neural network technology. By deeply analyzing the App code and extracting features, it automatically selects suitable methods for encryption, obfuscation, and compression, increasing the difficulty of reverse engineering the code, effectively preventing the App from being cracked, copied, and repackaged by attackers. Additionally, the use of encryption and obfuscation engines further enhances the security of the App.

Dingxiang iOS reinforcement has the following technical advantages:

- **Deep Analysis and Feature Extraction**: By thoroughly scanning and analyzing the application code, extracting key features, and automatically selecting the optimal reinforcement scheme.
- **Multi-layer Encryption and Obfuscation**: Combining multi-layer code encryption and obfuscation techniques to increase code complexity and prevent reverse engineering and cracking.
- **Compression Technology**: Optimizing code size while ensuring code security, improving application runtime efficiency.
- **Real-time Protection**: Using dynamic monitoring and real-time protection mechanisms to timely detect and prevent potential security threats.

These technical advantages ensure that Hanxiaocircle App can maintain a high level of security and stability in the face of complex and ever-changing security threats. 

Hanxiaocircle App is committed to becoming a trusted comprehensive platform for Korean dramas and entertainment for users, providing a safe, convenient, and rich entertainment experience for Korean drama enthusiasts.

Security-Focused Hanxiaocircle App Adopts Dingxiang iOS Reinforcement

Among the top ten high-frequency telecom network fraud cases disclosed by the Ministry of Public Security, a case involving the use of “screen sharing” technology to commit fraud attracted significant attention.

In June, the Ministry of Public Security released the top ten high-frequency telecom network fraud types. According to the report, in 2023, the average age of telecom network fraud victims was 37 years old, with 62.1% between 18 and 40 years old and 33.1% between 41 and 65 years old.

Among the top ten high-frequency telecom network fraud cases disclosed by the Ministry of Public Security, a case involving the use of “screen sharing” technology to commit fraud attracted significant attention. In May 2024, Ms. Du from Wuxi, Jiangsu, received a video call at home from a person claiming to be a criminal investigation team officer of the Wuxi Public Security Bureau. In the video, a fake “police officer” dressed in uniform told Ms. Du that her bank card was suspected of money laundering and needed her cooperation for investigation. Following the instructions, Ms. Du downloaded a conferencing software for screen sharing to cooperate with the “officer” in verifying the funds in her bank account. The “officer” said that Ms. Du needed to transfer the funds in her bank account to a designated “safe account” to prove her innocence. During this process, to demonstrate the normal flow of funds, the “officer” also instructed Ms. Du to take out a bank loan of 150,000 yuan and transfer it to the “safe account.” Ms. Du only realized she was being scammed when her family discovered the situation.

![2027072302.png](https://cdn.aisecurius.com/images/998/998e04d0-9ea7-421f-8f6d-3ed25ef430da.png)

In this case, the scammer induced the victim to enable screen sharing on her phone, leading to the leakage of her account passwords. Even if the scammer did not actively ask for information, they could see everything on the victim’s phone, including characters typed during password entry, received verification codes, and more, thereby transferring funds from the victim’s card.


#  How Scammers Use "Screen Sharing" for Fraud


"Screen sharing" is one of the application scenarios of Internet technology, where the screen information of one device is projected in real-time onto another device over the network. This includes screen apps, pop-up SMS, WeChat messages, other app notifications, the process of entering account passwords, unlocking, and all other real-time operations displayed on the screen. Originally designed to facilitate user sharing, this function has become a means for fraudsters to profit.

According to the analysis by Dingxiang Security Cloud Business Intelligence Center, scammers use “screen sharing” for fraud in four main steps:

1. **Information Gathering and Trust Building:** Scammers use illegal channels to obtain victim information and impersonate public security officers, shopping platform customer service, or campus loan cancellation staff to accurately match the victim’s situation and gain trust.

2. **Tricking Victims into Enabling "Screen Sharing":** Using apps with built-in screen sharing functions (QQ, Tencent Meeting, DingTalk, etc.), scammers can monitor all operations on the victim’s phone in real-time.

![2024071901.png](https://cdn.aisecurius.com/images/94e/94ec54a1-6b3e-4d4a-8d34-81d4fd333514.png)

3. **Monitoring Online Banking Accounts and Verification Codes:** Scammers instruct victims to download apps and enable screen sharing. Using the previously obtained bank card numbers, they log into online banking to perform transfer operations while monitoring the received SMS verification codes in real-time.
![2024071902.png](https://cdn.aisecurius.com/images/2d3/2d3c59f0-a9c9-44cd-89d8-29599ff15fd6.png)
4. **Stealing Funds:** Initially, scammers trick victims into transferring funds themselves. If the victim realizes the scam or hesitates, scammers use intercepted verification codes to directly perform transfers, moving the funds without the victim’s knowledge.


# Dingxiang Device Fingerprinting Can Prevent This Risk


On Android devices, “screen sharing” is implemented through the MediaProjection service. MediaProjection is a technology available in Android 5.0 and above, allowing developers to capture or record the screen. When the MediaProjection service is requested, the Android system pops up a permission request box. Only after the user clicks "Confirm" can screen sharing or recording begin.

![2024071903.png](https://cdn.aisecurius.com/images/7f7/7f74b560-d6c2-4c53-b7b5-21ea50d4fabd.png)

Dingxiang Device Fingerprinting can detect screen sharing calls from apps like DingTalk and Tencent Meeting in real-time and immediately issue alerts to warn users of potential risks.

When MediaProjection is initiated, the system automatically creates a virtual display. Dingxiang Device Fingerprinting can detect this additional display, confirm that the device is in “screen sharing” mode, and immediately report it to the risk control center. The risk control center will then issue an alert, warning the operator of the potential risk and advising caution.

Dingxiang Device Fingerprinting involves collecting and analyzing a device's hardware, software, and behavioral data to uniquely identify and recognize each device. By using Device Fingerprinting, virtual machines, proxy servers, simulators, and other maliciously controlled devices can be identified. This technology can analyze whether a device has multiple account logins, frequently changes IP addresses, or frequently changes device attributes, identifying abnormal or unusual user behaviors. This helps track and identify fraudulent activities, preventing insurance fraud incidents from occurring. Additionally, Device Fingerprinting can serve as an extra layer of identity verification, enhancing security during user login and transaction processes. By recording and comparing Device Fingerprinting, legitimate users and potential fraudsters can be distinguished. This technology supports Android, iOS, H5, public accounts, and mini-programs, effectively detecting risks such as simulators, rooted devices, jailbreaks, and injection hijacking. It features rapid countermeasures, efficient risk identification, over 99% stability, and 100% uniqueness.


# How Users Can Prevent "Screen Sharing" Risks


"Screen sharing" is increasingly being used in various forms of scams. The final step in these scams is to trick the victim into using or downloading network video conferencing software that provides "screen sharing" functionality. The goal is to spy on the user’s phone operations to obtain critical information such as bank account passwords.

1. **Do not casually enable the "screen sharing" function on your Android phone unless confirmed.**

2. **Immediately close any "screen sharing" invitations from strangers. Do not use the "screen sharing" function with strangers, as they can see all your operations, including entering passwords.**

3. **Be extremely cautious when dealing with private information, especially bank card passwords and verification codes. Do not disclose any such information to strangers.**

When receiving suspicious video calls or phone calls, stay calm. Use an excuse like a bad signal to hang up and then immediately call back through another method to confirm the legitimacy of the call, avoiding direct responses to potential scam content.

The Ministry of Public Security Disclosed This Telecom Fraud Case, Dingxiang Device Fingerprinting Can Prevent It

However, employment service websites face issues such as data leakage, identity theft, phishing attacks, and account security problems

Employment service websites play a crucial role in promoting local enterprise recruitment and talent employment. For enterprises, these websites provide a convenient and efficient recruitment channel, reducing recruitment costs, lowering advertising expenses, and simplifying recruitment processes. For job seekers, these websites serve as a two-way platform for self-presentation, with transparent job information and feedback 
mechanisms that enhance information transparency and optimize matching mechanisms. This increases supply-demand matching efficiency and simplifies recruitment and job search processes. Additionally, employment service websites offer services such as contract termination, dispatch, and redeployment, facilitating various employment relationships. In summary, employment service websites enhance local enterprise recruitment efficiency and talent supply-demand matching, becoming a vital force in driving local economic development and social stability.

![2024071708.png](https://cdn.aisecurius.com/images/b66/b6676486-5941-403b-8075-aae3ac311939.png)

However, employment service websites face issues such as data leakage, identity theft, phishing attacks, and account security problems. To effectively address these risks, Turpan Employment Service Network has strengthened data encryption and security protection, improved identity verification mechanisms to prevent identity theft, and chosen Dingxiang atbCAPTCHA as a tool for user registration and login identity verification.

Dingxiang atbCAPTCHA, based on AIGC technology, can prevent AI brute force attacks, automated attacks, and phishing threats, effectively preventing unauthorized access, account theft, and malicious operations, thus protecting system stability. It integrates 13 types of verification methods and various defense strategies, encompassing 4380 risk strategies, 112 types of risk intelligence, covering 24 industries and 118 types of risks. Its defense accuracy rate is as high as 99.9%, and it can quickly transform from risk to intelligence. Additionally, it supports seamless passing for secure users, reducing real-time countermeasure handling time to within 60 seconds, further enhancing the convenience and efficiency of digital login services, effectively safeguarding user accounts and website information security.

Turpan Employment Service Network is hosted by the Turpan City Human Resources and Social Security Bureau and organized by the City Public Employment Service Bureau. It is an employment information portal that integrates job seekers, employers, and employment services. As the most authoritative public employment service website in Turpan, the information published online is characterized by "authenticity, authority, and convenience." The website offers sections on talent recruitment, employment consultation, policies and regulations, employment promotion, job search and recruitment, career intermediaries, and self-employment. In addition to publishing job seekers' personal information and employer needs, it also undertakes multiple services such as online job fairs, signing, contract termination, dispatch, redeployment, reporting, file inquiry, and statistics between graduates and employers.

 Dingxiang Captcha Enhances Security for Turpan Talent Service Platform

Using forged voices, attackers utilize various media such as social media, email, remote meetings, online recruitment, and news information to target and attack their audience.

Voice cloning technology uses AI and algorithms to replicate a person's voice, making it sound almost identical to the original. This technology can be applied in movie dubbing and speech synthesis, and it has broad application prospects in fields such as healthcare, education, and virtual assistants. For example, it helps people who have lost their voice regain their speech ability, create audiobooks, create personalized digital assistants, and provide natural text-to-speech and speech translation services.

![2024071002.png](https://cdn.aisecurius.com/images/8fd/8fd4b464-761e-4c5d-8c3b-5b6d9d986873.png)

There are three main ways of voice cloning: playback-based cloning, synthesis-based cloning, and mimicry-based cloning.

1. **Playback-based Cloning**
This method involves copying the recording of the speaker's voice and then using playback and deep learning algorithms to simulate and replicate the speaker's voice. There are two main types: far-field detection and cut-and-paste detection. Far-field detection uses a hands-free phone to play the victim's recording, while cut-and-paste forges sentences requested from text-dependent systems.

2. **Synthesis-based Cloning**
Voice synthesis-based voice cloning uses software or hardware systems to convert text into natural speech in real-time. Such systems typically consist of three modules: a text analysis model, an acoustic model, and a vocoder. First, clear and well-structured original audio and corresponding text transcriptions need to be collected. Then, using this data, a text-to-speech model is trained to build a synthetic audio generation model. The text analysis module converts the input text into linguistic features, the acoustic module extracts the target speaker's parameters based on these features, and finally, the vocoder creates the human voice waveform based on these parameters, generating the final audio file.

3. **Mimicry-based Cloning**
Mimicry-based voice cloning, also known as voice conversion, is a method of converting the original speech of one speaker to sound like another speaker. This technology tries to mimic the target voice without changing the linguistic information by altering the style, tone, or prosody of the voice signal. This technique usually uses generative adversarial networks (GANs) for mimicry generation, offering high flexibility and quality, and can change the style, tone, or prosody of the voice while keeping the input text unchanged.


# Voice Cloning as a New Means for Scammers


The development of voice cloning technology has brought revolutionary changes to speech synthesis, personalized services, and other fields. However, the double-edged nature of technology has also led to an increase in fraud activities based on voice cloning. Voice cloning is being used by criminals to carry out various and increasingly rampant fraud schemes.

![2024071003.png](https://cdn.aisecurius.com/images/249/249445af-4757-47af-b07a-957a6739bc28.png)

**Social Media Fraud.** Scammers use cloned voices to impersonate public figures or the victim's relatives and friends on social media to carry out fraudulent activities.

**Social Tools or Phone Fraud.** By sending forged voice messages through social tools or directly making phone calls, scammers induce victims to disclose sensitive information or directly transfer money.

**Remote Meeting Attacks.** In remote meetings, scammers use cloned voices of participants to interfere or mislead, stealing business secrets or personal data.

**Fake News Information.** Scammers use cloned voices in fake news information to increase the credibility of the information, tricking victims into falling for it.


#  How to Prevent Voice Cloning Fraud


Using forged voices, attackers utilize various media such as social media, email, remote meetings, online recruitment, and news information to target and attack their audience. To prevent and combat voice cloning fraud, it is crucial to effectively identify and detect forged voices while also preventing the utilization and spread of voice cloning. Therefore, individuals should reduce sharing sensitive information and implement multiple preventive measures. Platforms utilized by attackers need to strengthen the identification of voice cloning and reduce fraudulent activities by analyzing and verifying user behavior and identity. Additionally, AI tools for voice anti-counterfeiting can provide security for voice communication.


###  Individual Preventive Strategies


1. **Reduce Sharing of Sensitive Information.** Avoid sharing sensitive information such as personal photos, voices, and videos on social media, and reduce public disclosure of private information such as personal accounts, family, and work to lower the risk of identity forgery. If "deepfake" content is discovered, it should be reported to social media administrators and law enforcement authorities immediately, and measures should be taken to delete and trace the source.

2. **Proactive Prevention by Voice Cloning Victims.** Set a "safety word" or "challenge question" known only to close friends, family, and colleagues, to verify the identity of the caller when receiving suspicious calls or messages. If the other party cannot provide the correct "safety word" or avoids the question, hang up immediately and confirm through known safe contact methods.

3. **Recipient Vigilance.** When receiving suspicious videos or calls, remain calm, make an excuse about bad signal to hang up, and call back immediately through other means to avoid directly responding to potential fraudulent content.



### Platform and Technology Preventive Measures



 **1.Enhance Voice Cloning Detection Technology.** Social media and communication platforms should strengthen their ability to identify voice cloning using advanced AI technology for voice authenticity verification, promptly detecting and intercepting suspicious content. Develop and apply AI tools for voice anti-counterfeiting, such as using voiceprint recognition technology to confirm the authenticity of the voice, providing security for voice communication.

2. **Identify Attackers Based on User Behavior and Identity.** Platforms should establish a security alert mechanism by analyzing user behavior patterns and identity information, monitoring and restricting suspicious activities such as abnormal logins and high-frequency message sending. Additionally, analyzing user behavior patterns such as mouse movement and typing style to identify anomalies, marking suspicious activities that deviate from normal use. Implement additional identity and device verification, and use large models to quickly screen vast amounts of data and detect subtle inconsistencies usually unnoticed by humans, identifying attackers' abnormal operations.

Dingxiang Device Fingerprinting can generate a unified and unique device fingerprint for each device. It builds a multi-dimensional recognition strategy model based on devices, environments, and behaviors to identify risk devices such as virtual machines, proxy servers, and emulators being maliciously controlled. It analyzes whether the device has multiple account logins, frequent IP address changes, or frequent device attribute changes, tracking and identifying fraudster activities. This helps enterprises achieve unified ID operations across all scenarios and channels, supporting cross-channel risk identification and control.

Dingxiang atbCAPTCHA, based on AIGC technology, can prevent threats such as AI brute force attacks, automated attacks, and phishing attacks, effectively preventing unauthorized access, account hijacking, and malicious operations, thereby protecting system stability. It integrates 13 verification methods and multiple prevention and control strategies, with 4,380 risk strategies, 112 types of risk intelligence, covering 24 industries and 118 types of risks. Its prevention and control accuracy rate is as high as 99.9%, and it can quickly convert risks into intelligence. It also supports seamless user experience for secure users, reducing real-time response handling to within 60 seconds, further enhancing the convenience and efficiency of digital login services.

Dingxiang Dinsight helps enterprises with risk assessment, anti-fraud analysis, and real-time monitoring, improving the efficiency and accuracy of risk control. Dinsight's average processing speed for daily risk control strategies is within 100 milliseconds, supporting the configuration and accumulation of multi-source data. It leverages experience in mature indicators, strategies, and models, as well as deep learning technology, to achieve self-monitoring and self-iterating mechanisms for risk control performance. Paired with the Xintell intelligent model platform, it can automatically optimize security strategies for known risks, detect potential risks based on risk control logs and data mining, and support risk control strategies for different scenarios with one-click configuration. Its technology, based on association networks and deep learning, standardizes the complex processes of data processing, mining, and machine learning, offering end-to-end modeling services from data processing, feature derivation, model building to final model deployment.


### Application of AI Tools for Voice Anti-counterfeiting


AI voice integration tools can also embed indelible watermarks, such as slight disturbances, random noise, or fixed background rhythms, which can be identified by the listener. Additionally, hardware for recording audio with AI voice integration tools can include built-in sensors to detect and measure biological signals produced by the human body while speaking, such as heartbeat, lung movements, vocal cord vibrations, and movements of the lips, jaw, and tongue. These recorded signals can be attached to the audio, providing verifiable information to the listener to distinguish whether the voice is naturally recorded or AI-forged.

How to Identify: Is the Voice on the Phone Cloned by Scammers?

On April 10, 2024, password management tool company LastPass disclosed through its official blog a scam involving the impersonation of its CEO

On April 10, 2024, password management tool company LastPass disclosed through its official blog a scam involving the impersonation of its CEO, Karim Toubba, using Deepfake technology, warning the public to beware of AI-generated fraudulent calls.

![2024070401.png](https://cdn.aisecurius.com/images/339/339dccff-e0fc-4cb2-a61c-26c595472459.png)

According to the blog, a fraudster impersonated LastPass CEO Karim Toubba by creating a WhatsApp account. The scammer then sent a series of messages to LastPass employees, including AI-generated voice messages imitating Karim Toubba's voice, multiple unanswered audio calls, and other communications. This created a sense of urgency to force employees into believing it was a genuine call from CEO Karim Toubba. Social engineering tactics in cybercrime often rely on creating urgency or panic to compel victims into hasty decisions.

However, the employee quickly identified this as a fraudulent call. WhatsApp is not a commonly used communication tool within LastPass, and the content of the voice messages exceeded the employee's expectations of normalcy. The employee promptly reported the incident to LastPass' security department. Subsequently, LastPass issued a notice confirming the attack did not compromise its overall security posture.

This attempted fraud highlights how advancements in AI technology, specifically voice cloning, are making cybercrime scams increasingly sophisticated. Furthermore, it underscores the importance of employee awareness training in mitigating such attacks.

These incidents continue to demonstrate the evolving complexity of cyberattacks. On the other hand, LastPass emphasizes the importance of employee awareness training in mitigating such attacks. Social engineering strategies often rely on creating urgency or panic to compel victims into hasty decisions.

AI voice cloning can generate synthetic voices for individuals with speech impairments, but it also serves as a tool for malicious actors seeking to profit from deception. To defend against AI voice cloning fraud, the Dingxiang Cloud Business Security Intelligence Center recommends several defensive measures:

1. **Multi-dimensional Voice Collection**: Enhance voice capture with additional dimensions. Hardware used for audio recording can include built-in sensors to detect and measure biological signals such as heartbeat, lung movements, vocal cord vibrations, and movements of the lips, jaw, and tongue. Recorded voices can be augmented with this information, providing verifiable details to distinguish natural recordings from AI fabrications.

2. **Embedded Indelible Watermarks in AI Voice Fusion Tools**: AI voice synthesis tools should incorporate indelible voice watermarks, such as subtle disturbances, random noise, or fixed background rhythms. These enable listeners to identify discrepancies.

3. **Establishment of "Safety Words"**: Pre-establish a "safety word" known to family, friends, colleagues, and partners. If receiving a suspicious call, request the caller to provide this "safety word." Failure to do so can promptly identify fraudulent calls and prevent panic.

4. **Hang Up and Redial**: If receiving a suspicious call, simulate poor signal and promptly hang up. Then redial the purported caller's number to confirm authenticity.

In conclusion, when confronted with AI voice cloning fraud, maintaining composure and vigilance is crucial alongside implementing effective prevention and response measures. Timely reporting and cooperation with law enforcement are essential in combating such criminal activities.


Renowned US Password Tool Company Falls Victim to AI Voice Cloning Scam

Liaoning Internet Security Department received reports from the public that their bank accounts were missing three thousand yuan, with several records showing transfers to unfamiliar accounts.


In May 2024, Liaoning Internet Security Department received reports from the public that their bank accounts were missing three thousand yuan, with several records showing transfers to unfamiliar accounts. Upon receiving the clues, Liaoning cyber police promptly launched an investigation. During the investigation, the Internet security department received similar reports multiple times, where money was sporadically transferred to unfamiliar accounts, mostly in amounts below 50,000 yuan. These cases shared the common traits of small transfer amounts and covert methods, raising high alert levels among law enforcement.

![2024070203.png](https://cdn.aisecurius.com/images/d05/d0580d7f-06c7-4d88-b4e9-4786fdd83665.png)

After several months of investigation, Liaoning cyber police uncovered an organized criminal group supported by AI technology. The criminal group, led by Ren Moumou, consisted of four "information officers" and "liaison officers" as key members, with an additional 27 "support personnel" providing assistance. They had clear divisions of responsibility: "information officers" Zhao Moumou and Liu Moumou were responsible for purchasing citizens' personal information elements—facial photos, mobile numbers, and bank card numbers. "Liaison officer" Liu Moumou was responsible for contacting overseas criminals, with "Qiu Mou" selling these elements. "Qiu Mou" then used AI software to create facial recognition verification videos using face images, subsequently stealing depositors' bank balances.


# The theft process of the criminal group


Dingxiang Defense Cloud Business Security Intelligence Center released a special issue titled "AI Face-swapping Threat Research and Security Strategy," revealing that the criminal group used AI face-swapping technology. Through deep learning algorithms, they synthesized the facial features of the target individuals into videos to bypass financial institutions' facial recognition verification systems. This application of technology enabled criminals to remotely control bank accounts without the depositors' awareness.

For financial institutions, the new type of telecom fraud based on "AI face-swapping" is particularly concerning. Fraudsters fabricate information, voice, video, and images, combining real and false identity information to create entirely new false identities for opening bank accounts or engaging in fraudulent purchases. Moreover, fraudsters can use "AI face-swapping" technology to quickly learn different banking processes and swiftly perpetrate fraud across multiple banks simultaneously.

![2024062703.jpg](https://cdn.aisecurius.com/images/727/727f6a32-2069-4abc-a621-3c159014cd8c.jpg)

Using a specific case intercepted by Dingxiang Defense Cloud Business Security Intelligence Center as an example of a new type of telecom fraud based on "AI face-swapping," the fraud process of the fraudster primarily involves four stages. "AI face-swapping" technology is just one critical factor in the fraud process; in other stages, if victims cannot discern and judge, they can easily step into traps following the instructions of fraudsters.

### 
1. Initially, the criminal group acquires large 
amounts of citizens' personal information, including facial photos, mobile numbers, and bank card numbers, through the black market or illegal means. This information may stem from various illegal database leaks or organized cyber attacks.


### 2. Utilizing AI technology
, the criminals use acquired facial photos as base material to generate high-quality fake faces through face synthesis algorithms. These faces may not exist in reality but closely resemble the original photos visually, sufficient to deceive most facial recognition systems.


### 3. Subsequently
, the criminal group uses AI software to embed the generated fake faces into videos, creating seemingly authentic facial recognition verification videos. These videos typically include fake faces, corresponding mobile numbers, and bank card information. They simulate real face-to-face verification processes, deceiving most facial recognition systems. Once the fake verification videos are created, the criminal group combines them with stolen bank card information.


### 4. The criminal group uses smartphones 
or computers through banking applications or online banking to transfer funds from victims' accounts to specified unfamiliar accounts, often during late nights or when users infrequently use their bank accounts, to avoid detection by victims.


# Defense: Multiple Technological Measures


[To combat "AI face-swapping" new telecom fraud, it is essential to verify offline when suspicious online incidents arise, increase communication ](https://www.aisecurius.com/blog/article/55)time, and employ exploratory measures such as requesting specific actions to expose potential flaws. Furthermore, enterprises are advised to adopt multiple technologies and methods. Additionally, guiding the positive application of AI technology and aggressively combating criminal activities are fundamental solutions.


### 1. Recognition of fraudulent "AI face-swapping" videos:
 During video chats, requesting actions like touching the nose or face to observe facial changes can help detect authenticity. A real person's nose will deform when pressed. Alternatively, requesting actions such as eating or drinking to observe facial changes, or asking to perform unusual gestures or expressions like waving or difficult hand gestures can aid in distinguishing between genuine and fake. Waving can cause facial data disturbance, resulting in slight shaking, flickering, or other abnormal conditions.


### 2. Comparison and identification of device information,
 geographical location, and behavioral operations: Dingxiang Device Fingerprinting identifies legitimate users and potential fraudulent behavior by recording and comparing device fingerprints. This technology uniquely identifies and recognizes each device, detecting manipulated devices such as virtual machines, proxy servers, or emulators. It analyzes behaviors like multiple account logins, frequent IP address changes, or frequent device attribute changes that deviate from typical user habits, assisting in tracking and identifying fraudster activities.


### 3. Identification of abnormal account behaviors: 
Activities such as remote logins, device changes, phone number changes, or sudden activity in dormant accounts require enhanced verification. Continuous identity verification during sessions is crucial to ensure consistent user identity throughout usage. Dingxiang atbCAPTCHA accurately distinguishes between human operators and machines, precisely identifying fraudulent behaviors, and real-time monitoring and intercepting abnormal activities.


### 4. Prevention of fake videos and images from "AI face-swapping"
: Dingxiang's comprehensive facial security threat perception solution intelligently verifies multiple dimensions of information such as device environment, facial information, image authenticity, user behavior, and interaction status. It quickly identifies and blocks injected attacks, live forgery, image forgery, camera hijacking, debugging risks, memory tampering, root/jailbreak attempts, malicious ROMs, emulators, and over 30 types of malicious behaviors. It can dynamically configure video verification strength and user-friendliness, implementing a dynamic mechanism of stronger verification for abnormal users while facilitating smoother access for legitimate users.


### 5. Uncovering potential fraud threats:
 Dingxiang Dinsight assists enterprises in risk assessment, anti-fraud analysis, and real-time monitoring to enhance the efficiency and accuracy of risk control. Dinsight's average daily processing speed for risk control strategies is within 100 milliseconds, supporting configurable access and deposition of multi-source data. Leveraging mature indicators, strategies, models, and deep learning technologies enables self-monitoring and iterative mechanisms for risk control. Paired with the Xintell intelligent model platform, it automatically optimizes security strategies based on risk logs and data mining of potential risks, supporting risk control strategies for various scenarios. Standardizing complex data processing, mining, and machine learning processes using correlation networks and deep learning technology provides end-to-end modeling services from data processing, feature derivation, model construction to final model deployment.



Stolen Bank Account: Tens of thousands quietly transferred away late at night

'Jihe Network,' digitizes linear textual content of ancient books. It collaborates with over 20 publishing institutions and research teams,

Ancient books, as precious heritage of human civilization, play an indispensable role in cultural and social development. They not only record and preserve ancient knowledge and wisdom but also serve as cornerstones for academic research, providing rich resources for understanding history and culture. The ideas and viewpoints within ancient books transcend time and space, stimulating modern thinking and innovation, and fostering exchange between different cultures. Moreover, ancient books constitute a significant part of cultural heritage, and through translation and digitization efforts, they can be better protected and disseminated.

The national comprehensive platform for the collation and digitization of ancient books, 'Jihe Network,' digitizes linear textual content of ancient books. It collaborates with over 20 publishing institutions and research teams, establishing fair trade rules and pricing systems to facilitate the circulation and dissemination of digital resources of ancient books.
![2024070202.png](https://cdn.aisecurius.com/images/306/3061f3d8-ed2f-406d-82a7-a7bdbd8d7234.png)
Jihe Network places great emphasis on data security and privacy protection, implementing a robust copyright protection mechanism to ensure the legality and compliance of every digital resource transaction. To prevent risks such as account verification through SMS being abused by malicious entities, Jihe Network has implemented a series of security measures and introduced Dingxiang seamless verification.


# How to Prevent Website SMS Abuse


Whether individuals face SMS bombardment or enterprises suffer SMS misuse, ultimately, it is due to attacks on the enterprise-side SMS interface. Effective security measures need to be taken to prevent the misuse of mobile SMS.


### 1. Set SMS sending time intervals:
 Set time intervals for repeated sending to the same number, generally set between 60-120 seconds, to prevent malicious attacks on the SMS interface.


### 2. Set SMS acquisition limits
: Restrict the number of times a particular phone number can receive SMS verification codes within a certain time period, such as different limits for 24 hours, 12 hours, 1 hour, etc.


### 3. Restrict SMS call IP addresses:
 Set the maximum sending volume for a single IP address within a certain time period, although this may cause false positives.
![2024070201.png](https://cdn.aisecurius.com/images/16f/16f02e3c-1b36-4398-94b2-6fe7d3af3b9a.png)

### 4. Deploy intelligent CAPTCHA:
 Before sending SMS verification codes, validate through CAPTCHA to prevent automated software sending. Dingxiang atbCAPTCHA provides additional security verification for SMS sending, preventing software-based sending, and using risk data and dedicated risk control models based on sender environment, devices, and more, it estimates possible abnormal operations for secure operations provide risk decision data.

Dingxiang atbCAPTCHA based on AIGC technology can prevent AI brute-force attacks, automated attacks, phishing attacks and other threats, effectively prevent unauthorized access, account theft and malicious operations, thus protecting system stability. It integrates 13 verification methods and multiple prevention and control strategies, compiling 4380 risk policies, 112 risk intelligence categories, covering 24 industries and 118 risk types. Its precision in prevention and control reaches 99.9%, and it can quickly convert from risk to intelligence. It also supports seamless secure user pass-through, with real-time handling capabilities reduced to within 60 seconds, further enhancing the convenience and efficiency of user login services.

Jihe Network is the first comprehensive platform for the collation and digitization of ancient books in China, developed and operated by Guolian (Beijing) Digital Media Technology Co., Ltd., a subsidiary of Zhonghua Book Company. Jihe Network integrates a variety of digital products of ancient books, providing unified user management and online purchasing services. It includes various digital products such as 'Chinese Classic Ancient Books Library,' 'Chinese Literature and History Academic Works Library,' 'Chinese Ancient Books Bibliography Database,' 'Chinese Stone Inscription Database,' 'Overseas Traditional Chinese Medicine Ancient Books Database,' 'Chinese Literature and History Reference Works Database,' 'Imperial Examination Database,' 'Chinese Woodblock New Year Prints Database,' 'Chinese Calligraphy Database,' 'Jihe Library,' 'Late Qing and Republic of China Literature Platform,' 'Southwest Associated University Special Database,' and more. The database types and themes are diverse, encompassing ancient literature, academic research, cultural arts, and thematic achievements. As of 2021, the 'Chinese Classic Ancient Books Library' has online resources exceeding 5,000 kinds, totaling approximately 2 billion characters, and has become one of the authoritative databases in its professional field."

Dingxiang atbCAPTCHA Helps Promote Research on Ancient Books and Culture

Mr. Li from Hebei was browsing a dating app when he was tempted and agreed to chat with the so-called "beauty" and downloaded an illegal chat software sent by the other party

As AI technology continues to develop, AI has become a new tool for online fraud. Criminals use deepfake technology to perform "AI voice impersonation" and "Deepfake", combining other people's faces and voices to create fake images and videos, creating fake identities, and impersonating others to induce victims to transfer money and other crimes. This type of fraud is varied and difficult to distinguish, and consumers are easily caught in scams. Without a doubt, in today's digital environment, AI has changed the nature of identity fraud.


# Four Types of People that Fraudsters Like to Impersonate


**1. Impersonating acquaintances**

[Ms. Li, a resident of Ordos City, Inner Mongolia, recently encountered a new type of telecommunications and network fraud case that used AI face-swapping technology.](https://www.aisecurius.com/blog/article/84) The fraudster impersonated Ms. Li's old classmate "Jia Mou" and contacted her through WeChat and QQ platforms. They used AI technology to create fake video calls and successfully gained Ms. Li's trust. Ms. Li received a WeChat friend request, and the other party's nickname and avatar were her old classmate "Jia Mou". After a brief identity confirmation via QQ video call, the fraudster asked Ms. Li for help with a cash flow problem and sent her a screenshot of a fake bank transfer record. Without verifying the recipient's information, Ms. Li transferred 400,000 yuan to the other party's account. When the other party continued to ask for money, Ms. Li realized that she might be being scammed, so she immediately contacted her real classmate and called the police. The Ordos police quickly activated the emergency stop payment mechanism and successfully intercepted and returned Ms. Li's 400,000 yuan.

![2024062501.png](https://cdn.aisecurius.com/images/22b/22b8287a-1b65-44fb-97dc-74fd64cea3a7.png)

**2. Impersonating beauties**

[Mr. Li from Hebei was browsing a dating app when he was tempted and agreed to chat with the so-called "beauty" and downloaded an illegal chat software sent by the other party. ](https://www.aisecurius.com/blog/article/84)However, the so-called "beauty" was actually a fake person created by the scammer. During the chat, the scammer took screenshots of Mr. Li's appearance and used "Deepfake" to create a pornographic video with Mr. Li's face. At the same time, the scammer also used the illegal chat software to obtain the mobile phone numbers of all the contacts in Mr. Li's mobile phone, and used this information to blackmail and extort Mr. Li. In order to get the scammer to delete the pornographic video, Mr. Li had no choice but to agree to transfer 120,000 yuan to the other party.

**3. Impersonating colleagues**

[In January of this year, there was a 200 million Hong Kong dollar "Deepfake" video conference fraud case in Hong Kong. The characteristic of this fraud case is that it uses AI deepfake technology. ](https://www.aisecurius.com/blog/article/50)The fraudster first collected facial and voice data of the target company's senior management personnel, and then used AI technology to "face swap" this data onto the scammer's body to create a seemingly real video conference, which led to the company employees participating in the video conference transferring 200 million Hong Kong dollars to the scammer's account.

![2024063704.png](https://cdn.aisecurius.com/images/0c7/0c71c6c2-5294-40da-8e7d-766f4b571f96.png)

**4. Impersonating family members**

[In December 2023, an overseas student was "kidnapped" abroad, and his parents were asked by the "kidnappers" for a ransom of 5 million yuan. ](https://www.aisecurius.com/blog/article/47)They also received a video of the "hostage" being controlled and harmed. Through on-site investigation and international police cooperation, 5 hours later, the student involved, Xiaoxia, was successfully rescued - not from the kidnappers, but at the immigration port of entry in the country where his school is located. The truth was also revealed: the scammer used deepfake technology to create Xiaoxia's voice and kidnapping video, and then sent it to Xiaoxia's family to defraud money.

![2024062707.png](https://cdn.aisecurius.com/images/31f/31f92432-e84e-4662-9394-ee7176914f06.png)

How victims' information is leaked

Before carrying out new telecommunications network frauds such as "AI voice imitation" and "Deepfake," fraudsters first collect targets' information, including portraits, contact details, home addresses, work information, lifestyle details, as well as photos, videos, and voice recordings of victims. The avenues through which this information is obtained are diverse and include:


### 1. Social media
: Fraudsters can gather victims' photos and personal information through social media platforms and other publicly available online resources. Many individuals share a substantial amount of personal photos and videos on social media, which can be exploited by malicious actors. Additionally, some forums and online platforms may publicly display users' photos and videos, allowing fraudsters to gather relevant information.


### 2. Network data breaches
: Large-scale network data breaches can lead to the exposure of personal information, including photos, videos, and voice recordings. Fraudsters can access this leaked data through dark web platforms to execute fraudulent activities.


### 3. Phishing and malware
: Fraudsters use phishing emails or malicious software to obtain victims' personal information, including photos and videos. Once they gain access to the victim's device, they can further acquire more personal data. For example, on February 15, [2024, the foreign security company Group-IB announced the discovery of malicious software named "GoldPickaxe." The iOS version of this malware deceives users into performing facial recognition, submitting identification documents, and then uses the user's facial information for "Deepfake" fraud.](https://www.aisecurius.com/blog/article/51)


### 4. Public events and on-site activities
: In public places or specific events such as conferences, exhibitions, and social gatherings, fraudsters may collect victims' photos and videos.

Among these methods, collecting a large number of photos, information, and videos from publicly accessible social media is prevalent. Since social media has become an indispensable part of daily life, it is crucial to avoid oversharing sensitive information on social media platforms, as it serves as a warehouse for fraudsters to obtain materials for fraud.

Especially, every dynamic update and every photo shared by victims on social media contain a wealth of sensitive information, which has become a repository for fraudsters to obtain materials for fraud. Based on the collected sensitive information, fraudsters perform feature extraction, enabling them to create fake facial videos. By extracting voice samples from video clips on social media, with samples as short as 30 seconds to 1 minute, fraudsters can produce highly realistic voice clones.


# Technology Prevention: New Telecommunications Network Fraud


[Dingxiang Defense Cloud Business Security Intelligence Center has released a special report titled "Deepfake Threat Research and Security Strategies,"](https://www.aisecurius.com/blog/article/55) systematically detailing the industrial chain behind new telecommunications network fraud based on "Deepfake."

To prevent "Deepfake" new telecommunications network fraud, it is necessary to verify offline when suspicious points are discovered online, increase communication time, and adopt probing measures, such as requesting the other party to perform specific actions, to expose potential loopholes. Additionally, it is recommended that enterprises employ multiple technologies and methods. Furthermore, promoting the positive application of AI technology and rigorously cracking down on criminal activities are fundamental solution
### 1. Identification of "Deepfake" fraudulent videos: 
deos: During video chats, one can request the other party to press their nose or face to observe facial changes; pressing a real nose will cause deformation. Alternatively, requesting actions like eating or drinking to observe facial changes can be effective. Asking for unusual gestures or expressions, such as waving or difficult hand gestures, can also help distinguish authenticity. During waving, facial data interference occurs, causing slight tremors, flickers, or abnormal situatio
### 
2. Comparison and identification of device information, geographical location, and behavioral operations
ations: Dingxiang Device Fingerprinting distinguishes between legitimate users and potential fraudulent behavior by recording and comparing Device Fingerprinting. This technology uniquely identifies and recognizes each device, detecting manipulated devices like virtual machines, proxy servers, and emulators. It analyzes behaviors such as multiple account logins, frequent IP address changes, and device attribute changes that deviate from user habits, aiding in tracking and identifying fraudulent activitie
### 3. Recognition of abnormal account activities: 
ties: Increased verification is required for activities such as remote logins, device changes, phone number changes, or suddenly active dormant accounts. Continuous identity verification during sessions is crucial to maintain consistent user identity. Dingxiang atbCAPTCHA accurately differentiates between human operators and machines, precisely identifying fraudulent behaviors, and provides real-time monitoring and interception of abnormal activities.

![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)


### 4. Prevention of fraudulent "Deepfake" videos and images
: Dingxiang's comprehensive panoramic facial security threat perception solution employs multi-dimensional verification of device environments, facial information, image authenticity, user behavior, and interaction status. It quickly identifies injected attacks, live body forgeries, image forgeries, camera hijacking, debugging risks, memory tampering, Root/Jailbreak, malicious ROMs, simulators, and over 30 other types of malicious behaviors. Upon detecting forged videos, fake facial images, or abnormal interaction behaviors, it automatically blocks operations. The solution offers flexible configuration of video verification strength and user friendliness, implementing a dynamic mechanism for enhanced verification of abnormal users while employing atbCAPTCHA for normal users.


### 5. Unearthing potential fraud threats:
 Dingxiang Dinsight aids enterprises in risk assessment, anti-fraud analysis, and real-time monitoring, enhancing risk control efficiency and accuracy. Dinsight's average processing speed for daily risk control strategies is within 100 milliseconds, supporting configuration-based access and deposition of multi-source data. Leveraging mature indicators, strategies, models, and deep learning technology, it achieves self-monitoring and self-iterative mechanisms for risk control performance. Paired with the Xintell intelligent model platform, it optimizes security strategies automatically for known risks and configures support for different scenarios based on risk control logs and potential risk mining. Standardizing complex data processing, mining, and machine learning processes based on associative networks and deep learning technology, it provides end-to-end modeling services from data processing, feature derivation, model construction to final model deployment.

New Types of Telecommunications and Network Fraud: Four Types of People that Fraudsters Like to Impersonate

Ms. Li, a resident of Ordos City, Inner Mongolia, recently fell victim to a new telecommunications fraud scheme. 

Ms. Li, a resident of Ordos City, Inner Mongolia, recently fell victim to a new telecommunications fraud scheme. Fraudsters utilized DEEPFAKE technology to synthesize a video call resembling her old classmate, which became a crucial tool in deceiving Ms. Li's trust.

![2024062501.png](https://cdn.aisecurius.com/images/22b/22b8287a-1b65-44fb-97dc-74fd64cea3a7.png)

The impostor pretended to be Ms. Li's old classmate "Jia," establishing contact via WeChat and QQ platforms and using DEEPFAKE technology to forge a video call, successfully gaining Ms. Li's trust.

Ms. Li received a friend request on WeChat from an account with the nickname and profile picture matching her old classmate "Jia." After briefly confirming the identity through a QQ video call, the fraudsters claimed they needed money urgently and asked Ms. Li to transfer funds, providing a fake bank transfer screenshot. Without verifying the receipt, Ms. Li transferred 400,000 RMB into the fraudster's account. When they demanded more transfers, Ms. Li realized she might be deceived, immediately contacting her real old classmate and reporting the incident to the police.

![2024062503.png](https://cdn.aisecurius.com/images/621/6211a325-5eda-45f1-8b32-a56a16094924.png)

The Ordos police swiftly activated emergency measures to freeze the transaction and successfully returned the 400,000 RMB to Ms. Li. Through investigation, it was found that the WeChat account in question was linked to a person named Liu, who had registered 12 WeChat accounts under his name, engaging in suspicious activities. Liu eventually admitted to renting out his WeChat account to the fraudsters after seeing an opportunity in a QQ part-time job group about "earning money by renting out verified social media accounts."
![2024062502.png](https://cdn.aisecurius.com/images/5f5/5f54afe7-136d-4ce9-9ecc-f8fd89bbdc7b.png)
Mr. Li from Hebei also encountered a DEEPFAKE-based new telecommunications fraud case, where fraudsters induced him to engage in explicit conversations via a dating app, leading to the download of illegal chat software. They used AI technology to fabricate a pornographic video and extorted 120,000 RMB from him.


# Frauds Utilizing DEEPFAKE Technology


These incidents underscore the application of DEEPFAKE technology in fraudulent activities and the murky gray industry behind it.

Firstly, fraudsters target individuals on dating platforms using tactics like explicit conversations to lure victims into downloading illegal chat software, thereby illegally obtaining their contact lists, photo libraries, and other private information.

Secondly, DEEPFAKE technology captures facial features of victims through software, combining AI synthesis techniques to create fabricated video content.

Lastly, fraudsters utilize custom apps to distribute these videos, carrying out extortion and fraud.

The new telecommunications fraud has a complete industrial chain: upstream involves illegally obtaining citizens' private information, midstream utilizes custom scripts for fraud, and downstream involves money laundering and cashing out. Citizens' personal information is priced and sold on the dark web, with illegal links, QR codes, and apps becoming primary means of information extraction.


# Technical Measures to Prevent New Telecommunications Fraud Using DEEPFAKE


[Dingxiang Defense Cloud Business Security Intelligence Center has released an Intelligence Special Issue titled "DEEPFAKE Threat Research and Security Strategies," systematically outlining the industrial chain behind new telecommunications fraud based on DEEPFAKE technology.](https://www.aisecurius.com/blog/article/55)

To prevent DEEPFAKE-based new telecommunications fraud, it is essential to detect suspicious activities online and verify offline, increase communication time, and take probing measures such as requesting specific actions from the other party to expose potential loopholes. Additionally, it is recommended that enterprises adopt multiple technologies and methods. Furthermore, promoting the positive application of AI technology and harshly cracking down on criminal activities are fundamental solutions.


## 1. Identification of DEEPFAKE fraudulent videos: 
During video chats, request the other party to press their nose or face to observe facial changes. A real person's nose will deform when pressed. Alternatively, ask them to eat or drink and observe facial changes. Requesting unusual actions or expressions, such as waving or performing difficult gestures, can also help distinguish authenticity. These actions may cause facial data interference, resulting in slight tremors, flickers, or other anomalies during waving.


## 2. Comparison and identification of device information, geolocation, and behavioral operations can detect and prevent abnormal activities
. Dingxiang Device Fingerprinting identifies legitimate users and potential fraudulent behaviors by recording and comparing device fingerprints. This technology uniquely identifies each device, distinguishing manipulated devices like virtual machines, proxy servers, and emulators. It analyzes behaviors such as multiple account logins, frequent IP address changes, and device attribute changes that deviate from normal user habits, aiding in tracking and identifying fraudsters' activities.


## 3. Enhanced verification for actions such as remote login from different locations, device changes, 
phone number changes, and sudden activity spikes is crucial. Continuous identity verification during sessions is vital to ensure consistency throughout user interactions. Dingxiang atbCAPTCHA accurately distinguishes between human operators and machines, precisely identifying fraudulent behaviors, and monitors and intercepts abnormal activities in real-time.

![2024062705.png](https://cdn.aisecurius.com/images/e06/e06b5696-4cde-4099-bdfe-b33389a1def8.png)
## 4. AI technology and human verification systems for facial anti-fraud are employed to prevent DEEPFAKE fake videos and images.
 Dingxiang's comprehensive facial security threat perception solution employs multi-dimensional verification based on device environment, facial information, image authentication, user behavior, and interaction status. It swiftly identifies various malicious activities such as injection attacks, live forgery, image forgery, camera hijacking, debugging risks, memory tampering, Root/jailbreak, malicious ROMs, and emulators, automatically blocking operations upon detecting fabricated videos, fake facial images, or anomalous interaction behaviors. The system can dynamically adjust video verification strength and user-friendliness, implementing a mechanism for enhanced verification of abnormal users while employing atbCAPTCHA for regular users.


## 5. Unearthing potential fraud threats and thwarting complex DEEPFAKE attacks is achieved through 
Dingxiang Dinsight, which aids enterprises in risk assessment, anti-fraud analysis, and real-time monitoring to enhance risk control efficiency and accuracy. Dinsight processes daily risk control strategies in less than 100 milliseconds, supports configurable access and deposition of multi-party data, and optimizes security policies automatically based on known risks using risk logs and data mining. The Xintell intelligent model platform, in conjunction with Dinsight, enables automatic optimization of security strategies for various scenarios based on mature indicators, strategies, models, and deep learning technology, standardizing complex data processing, mining, and machine learning processes from data handling and feature derivation to model construction and deployment.


Continuous Fraud Cases! Rampant New Telecommunications Fraud Based on DEEPFAKE Technology

Facing these challenges, e-commerce platforms and regulatory authorities need to take more effective measures. On one hand, 

This year's "618" results are impressive. Data shows that mainstream e-commerce platforms have achieved remarkable performances, reflecting not only consumers' trust and support for these platforms but also their comprehensive upgrades in marketing strategies, supply chain management, and logistics. Livestream e-commerce, as a bridge connecting consumers with brands and products, has increasingly highlighted its role, making information exchange between market supply and demand more transparent and efficient.

With the rapid development of the e-commerce industry, issues related to illicit and gray practices are becoming increasingly prominent. During large-scale promotional events like "618," e-commerce platforms have had to relax risk control rules moderately to ensure user experience and system stability, inadvertently providing opportunities for illicit activities. Exploiting new technologies and methods, illicit actors launch attacks in more covert and sophisticated ways, posing significant challenges to both e-commerce platforms and consumers.

![2024062001.jpg](https://cdn.aisecurius.com/images/07c/07c28720-c4b9-4401-a009-a7ba145c71de.jpg)

Facing these challenges, e-commerce platforms and regulatory authorities need to take more effective measures. On one hand, e-commerce platforms need to strengthen technical defenses, enhance risk control capabilities, and promptly detect and intercept fraudulent activities. On the other hand, regulatory authorities need to intensify efforts against illicit practices, enforce laws to punish illegal behaviors, and maintain market order.


# Woolfarming Hidden in Livestream Rooms


During a liquor e-commerce livestream on May 20th, the host achieved sales exceeding 100 million yuan in just 43 minutes, selling 30,000 bottles of liquor. However, these "flash sale" special-priced goods have become channels for some "woolfarmers" to mass purchase. They infiltrate livestream rooms, use illegal plug-in software to snatch "flash sale" and "limited purchase" products, and then resell them at inflated prices, profiting immensely.

These "flash sale" special-priced goods have become channels for some "woolfarmers" to mass purchase. They infiltrate livestream rooms, use illegal plug-in software to snatch "flash sale" and "limited purchase" products, and then resell them at inflated prices, profiting immensely. "Woolfarmers" in e-commerce livestream rooms illegally profit through technical means, causing significant losses not only to e-commerce platforms but also disrupting the shopping experiences of ordinary consumers.
![2024061403.jpg](https://cdn.aisecurius.com/images/a39/a39c2f17-0df4-476a-a382-eb2e5d084965.jpg)

During shopping festivals, countless consumers eagerly anticipate purchasing items, yet many discounted or special-priced products always appear as "sold out" or "temporarily out of stock." What's the reason behind this? It turns out many popular, limited-supply, or high-value-for-money items are snatched away by woolfarmers.

Woolfarmers engaging in activities like grabbing coupons and snatching special-priced goods not only harm the legitimate interests of consumers but also cause major economic losses to consumers, merchants, and e-commerce platforms. Statistics show that 70% to 80% of e-commerce enterprises' marketing expenses are consumed by woolfarmers, resulting in annual losses amounting to hundreds of billions of yuan.


# Merchant Order Fakery Misleading Consumers


"If you fake orders, the worst that could happen is getting caught and banned by the platform; however, if you don't fake orders, you'll likely be starved to death by other fake order sellers or major sellers on the platform, or even starved to death on the platform," said a merchant on a livestreaming platform.

Based on the rules of livestream e-commerce platforms, when buyers search for a product using keywords on the platform, the e-commerce platform calculates the weight of all pages related to that product and presents them to the buyer in order of relevance. The higher a product ranks, the higher its exposure and click-through rates, leading to higher potential orders. Conversely, products that rank lower are rarely seen by buyers and thus have minimal chances of making sales.

There are hundreds, even thousands of indicators that affect the ranking of a product, with the three most important indicators being sales volume, conversion rate, and positive review rate. While this algorithm saves buyers a lot of time in choosing and comparing products, it has triggered a vicious cycle among merchants where "products ranking higher sell better, while products ranking lower sell worse." Therefore, to improve product rankings, many merchants have started focusing on these three indicators, giving rise to the practice of "faking orders."

Analysis by Dingxiang Defense Cloud Business Security Intelligence Center has found that e-commerce order fakery mainly occurs through software-based and human-based methods. Those involved in executing order fakery tasks in the gray market can earn commissions from merchants and profit from selling order faking cheating software.


## Software-based fakery involves account manipulation;
 gray market entities have a large number of e-commerce platform accounts. These accounts are obtained through mass registration on one hand and through illegal means such as the dark web or illegal methods like database breaches or brute-force attacks on already registered user accounts.

Upon receiving orders from merchants, gray market entities use cheating software to fake orders. This software can handle all operations from registration to purchase, including mass registration, intelligent account management, bulk linking of delivery addresses, bulk linking of credit cards, enabling two-step verification, direct reviews, leaving reviews, liking, adding to wish lists, adding to cart, and clearing cart, etc. Depending on the merchant's requirements, cheating software can simulate human behaviors to the maximum extent possible by randomly setting browsing paths, clicking on advertisements, answering questions, clicking on images, etc.
![2024061204.jpg](https://cdn.aisecurius.com/images/6cc/6cc16e27-c900-4a6e-96c1-0e0a813ac17c.jpg)

Cheating software from the gray market supports multiple accounts logging into a single device, isolates cookie data, modifies browser fingerprints, and assigns a foreign proxy IP address based on the logged-in account, ensuring each faked order account operates within a virtual "independent environment."


## Human-based crowdsourcing involves mobilizing real accounts to fake orders,
 with accounts not controlled by the gray market but operated by real individuals.

After a merchant issues an order fakery request, gray market entities immediately send product links in "order faking groups" and arrange batches of accounts to buy the merchant's products. In the process of faking orders, specific operational standards are set for order faking accounts to circumvent e-commerce platform risk control monitoring. For example, they search for keywords provided by the merchant on the e-commerce platform, randomly browse any product for 3 to 5 minutes, then find the target store's product needed for the faked order based on the merchant's information. After opening the designated product page provided by the merchant, they continue to browse the product details for 3 to 5 minutes before proceeding to place the order. These operations make the faked order appear as if it were a genuine customer's actual purchase behavior of "comparing products from three stores."

According to the merchant's requirements for shopping reviews and confirmation, the merchant must pay a certain percentage of "commission" to the gray market and provide "rebates" to the faked order account. There are various ways to provide rebates. Some gray market entities require merchants to directly "refund" to the faked order account after confirming the order.


# Merchant Order Fakery Misleading Consumers


"If you fake orders, the worst that could happen is getting caught and banned by the platform; however, if you don't fake orders, you'll likely be starved to death by other fake order sellers or major sellers on the platform, or even starved to death on the platform," said a merchant on a livestreaming platform.

Based on the rules of livestream e-commerce platforms, when buyers search for a product using keywords on the platform, the e-commerce platform calculates the weight of all pages related to that product and presents them to the buyer in order of relevance. The higher a product ranks, the higher its exposure and click-through rates, leading to higher potential orders. Conversely, products that rank lower are rarely seen by buyers and thus have minimal chances of making sales.

![2024061404.jpg](https://cdn.aisecurius.com/images/554/554fa25e-1802-4b68-85a8-0d29cf11a1f0.jpg)

There are hundreds, even thousands of indicators that affect the ranking of a product, with the three most important indicators being sales volume, conversion rate, and positive review rate. While this algorithm saves buyers a lot of time in choosing and comparing products, it has triggered a vicious cycle among merchants where "products ranking higher sell better, while products ranking lower sell worse." Therefore, to improve product rankings, many merchants have started focusing on these three indicators, giving rise to the practice of "faking orders."

Analysis by Dingxiang Defense Cloud Business Security Intelligence Center has found that e-commerce order fakery mainly occurs through software-based and human-based methods. Those involved in executing order fakery tasks in the gray market can earn commissions from merchants and profit from selling order faking cheating software.

Software-based fakery involves account manipulation; gray market entities have a large number of e-commerce platform accounts. These accounts are obtained through mass registration on one hand and through illegal means such as the dark web or illegal methods like database breaches or brute-force attacks on already registered user accounts.

Upon receiving orders from merchants, gray market entities use cheating software to fake orders. This software can handle all operations from registration to purchase, including mass registration, intelligent account management, bulk linking of delivery addresses, bulk linking of credit cards, enabling two-step verification, direct reviews, leaving reviews, liking, adding to wish lists, adding to cart, and clearing cart, etc. Depending on the merchant's requirements, cheating software can simulate human behaviors to the maximum extent possible by randomly setting browsing paths, clicking on advertisements, answering questions, clicking on images, etc.

Cheating software from the gray market supports multiple accounts logging into a single device, isolates cookie data, modifies browser fingerprints, and assigns a foreign proxy IP address based on the logged-in account, ensuring each faked order account operates within a virtual "independent environment."

Human-based crowdsourcing involves mobilizing real accounts to fake orders, with accounts not controlled by the gray market but operated by real individuals.

After a merchant issues an order fakery request, gray market entities immediately send product links in "order faking groups" and arrange batches of accounts to buy the merchant's products. In the process of faking orders, specific operational standards are set for order faking accounts to circumvent e-commerce platform risk control monitoring. For example, they search for keywords provided by the merchant on the e-commerce platform, randomly browse any product for 3 to 5 minutes, then find the target store's product needed for the faked order based on the merchant's information. After opening the designated product page provided by the merchant, they continue to browse the product details for 3 to 5 minutes before proceeding to place the order. These operations make the faked order appear as if it were a genuine customer's actual purchase behavior of "comparing products from three stores."

According to the merchant's requirements for shopping reviews and confirmation, the merchant must pay a certain percentage of "commission" to the gray market and provide "rebates" to the faked order account. There are various ways to provide rebates. Some gray market entities require merchants to directly "refund" to the faked order account after confirming the order.


# Preventing Various Types of Fraud on Platforms


Based on the cheating and fraud in livestream rooms, Dingxiang Defense Cloud Business Security Intelligence Center recommends integrating multiple product combinations on the business side for further control to ensure the normal order of livestream platforms.

Ensuring Client-Side Security. Dingxiang's Adaptive App Shielding, based on graph neural network technology, deeply analyzes and extracts code features. It automatically selects suitable methods for obfuscation based on the characteristics of different code blocks, significantly increasing the difficulty of reverse analysis and effectively reducing computational performance consumption by 50%. Through encryption and obfuscation engines, it enhances the security of App code, effectively preventing attacks such as cracking, copying, and repackaging by attackers.

Identification of Fraudulent Accounts. Dingxiang atbCAPTCHA, powered by AIGC technology, prevents AI brute-force cracking, automated attacks, phishing attacks, and other threats. It effectively prevents unauthorized access, account hijacking, and malicious operations, thereby protecting system stability. It integrates 13 verification methods and multiple control strategies, aggregating 4380 risk policies, 112 risk categories, covering 24 industries and 118 risk types. Its precision in risk control reaches up to 99.9%, with rapid transformation from risk to intelligence. It supports seamless authentication for secure users and achieves real-time response and disposal within 60 seconds, further enhancing user login service experience in terms of convenience and efficiency.

Identification of Fraudulent Devices. Dingxiang Device Fingerprinting integrates internal information of multi-platform devices to generate unified and unique Device Fingerprinting for each device. It builds multidimensional identification strategy models based on device, environment, and behavior to detect risks such as virtual machines, proxy servers, and emulators manipulated maliciously. It analyzes anomalies such as multiple account logins, frequent IP address changes, and device attribute changes that do not align with user habits, tracks and identifies fraudulent activities, and helps enterprises achieve unified operation of the same ID across all scenarios and channels, facilitating cross-channel risk identification and control.

Uncovering Potential Fraud Threats. Dingxiang Dinsight assists enterprises in risk assessment, anti-fraud analysis, and real-time monitoring, enhancing the efficiency and accuracy of risk control. Dinsight processes daily risk control strategy at an average speed of less than 100 milliseconds, supporting configurable access and sedimentation of multi-party data. It leverages mature indicators, strategies, and model experience reserves, as well as deep learning technology, to achieve self-monitoring and self-iteration mechanisms for risk control. Paired with the Xintell intelligent model platform, it optimizes security strategies automatically for known risks and configures support for risk control strategies across different scenarios based on risk control logs and data mining of potential risks. It standardizes complex processes of data processing, mining, and machine learning based on associative networks and deep learning technology, providing a one-stop modeling service from data processing, feature derivation, model construction to final model deployment.


Inventory: Fraudulent Practices in Livestream E-commerce

Following the conclusion of the mid-year promotion "618," concerns about the false prosperity behind livestreaming e-commerce are gradually emerging. 

Following the conclusion of the mid-year promotion "618," concerns about the false prosperity behind livestreaming e-commerce are gradually emerging. Many consumers have reported that the "popular items" in livestream rooms often have questionable quality, with some rooms using tools to artificially inflate traffic and positive reviews, creating a false impression of prosperity that misleads consumers into making purchases.
![2024062001.jpg](https://cdn.aisecurius.com/images/07c/07c28720-c4b9-4401-a009-a7ba145c71de.jpg)


Liu Tian (pseudonym), a Beijing office worker, expressed doubts about the authenticity of positive reviews due to the repetitive nature of comments and lack of genuine interaction in livestream rooms. Meanwhile, Li Xia (pseudonym), a "mommy" from Shanghai, encountered significant discrepancies between clothing purchased from livestreams and their descriptions. Media investigations have uncovered instances of abnormal data spikes in certain livestream rooms, where viewership surged from 20 to tens of thousands within just five minutes, inundating the barrage with irrelevant information while genuine inquiries about the products were scant. Even more shocking was a beauty device that purportedly sold over ten thousand units during the livestream but plummeted to just over one thousand after it ended.

Insiders revealed that some top livestreamers spend hundreds of thousands to boost their popularity, and livestream rooms with return rates exceeding 60% likely engage in data falsification.

Livestream viewership boosting is widespread, primarily achieved through two methods: automated software (botting) and manual group control. Bots simulate app interactions via computer programs to create artificial popularity, but these methods are increasingly thwarted as e-commerce platforms enhance their defenses. Gray market actors have shifted to using group control devices, operating multiple smartphones simultaneously to automate likes, follows, scripted messages, and other actions to fabricate the illusion of thriving livestream rooms.

Media outlets have contacted a water army team that offers services such as inflating livestream viewership, interactive comments, livestream donations, and even inflating product sales figures. The team leader claimed their accounts can mimic real users, making them difficult to distinguish. To create authentic livestream interactions, the water army team engages in scripted interactions based on questions and phrases provided by the livestreamers. Additionally, they provide services for maliciously reporting other livestream rooms and fabricating orders to boost sales.

The State Administration for Market Regulation recently published the "Interim Provisions on the Prohibition of Unfair Competition on the Internet," Article 9 of which specifically prohibits operators from engaging in false transactions, fictitious transaction volumes, fabricated user reviews, and falsified traffic data to ensure consumers can obtain genuine and accurate product information.


# Livestream Platforms: How to Identify Cheating in View Count


In response to cheating in livestream view counts, Dingxiang Defense Cloud Business Security Intelligence Center recommends deploying a combination of products on the business side for further control, ensuring the normal operation of livestream platforms.


## Ensuring Client-Side Security.
 Dingxiang's adaptive App shielding, based on graph neural network technology, enhances code obfuscation by analyzing and extracting code features deeply. It automatically selects appropriate methods for obfuscation based on different code characteristics, significantly increasing the difficulty of reverse engineering and effectively reducing computational consumption by 50%. Through encryption and obfuscation engines, it enhances the security of App code, preventing attacks such as cracking, duplication, and repackaging.


## Identification of View Count Manipulation Accounts. 
Dingxiang's atbCAPTCHA, employing AIGC technology, prevents AI-driven brute force attacks, automated assaults, and phishing threats. It safeguards against unauthorized access, account hijacking, and malicious operations, thereby ensuring system stability. It integrates 13 verification methods and multiple control strategies, encompassing 4380 risk policies, 112 risk categories, covering 24 industries and 118 risk types. It achieves precision in risk control up to 99.9% and swiftly transforms risks into actionable intelligence. It supports seamless authentication for secure users and offers real-time response capabilities within 60 seconds, further enhancing user login service convenience and efficiency.


## Recognition of View Count Manipulation Devices. 
Dingxiang Device Fingerprinting establishes unified and unique device fingerprints by internally integrating multi-device information. It constructs multidimensional recognition strategies based on device, environment, and behavior to identify risks such as virtual machines, proxy servers, and emulators manipulated maliciously. It analyzes abnormal behaviors such as multiple account logins, frequent IP address changes, and device attribute alterations that deviate from user habits, tracks and identifies fraudulent activities, and aids enterprises in unified operations under the same ID across all scenarios and channels, facilitating cross-channel risk identification and management.


## Detection of Potential Fraud Threats.
 Dingxiang Dinsight real-time risk control engine assists enterprises in risk assessment, anti-fraud analysis, and real-time monitoring, enhancing the efficiency and accuracy of risk control. Dinsight achieves an average processing speed of daily risk control strategies within 100 milliseconds, supports configurable access and deposition of multi-source data, and leverages mature indicators, strategies, models, and deep learning technology for self-monitoring and iterative enhancement of risk control mechanisms. Combined with the Xintell intelligent model platform, it optimizes security strategies automatically for known risks, configures support for risk control strategies across different scenarios based on risk control logs and data mining of potential risks. Standardizing complex processes of data handling, feature derivation, model building, and final model deployment, it provides an integrated modeling service from data processing to model deployment.


After 618: Consumers realize that the popularity of livestreaming e-commerce is "boosted" artificially

E-commerce shopping festivals are not only a feast for consumers but also a honey pot for black and grey industries. On the one hand, during major events like 618

E-commerce shopping festivals are not only a feast for consumers but also a honey pot for black and grey industries. On the one hand, during major events like 618, many e-commerce platforms lower their business security sensitivity and relax some risk control rules to avoid impacting customer experience and ensure stable business operations. On the other hand, black and grey industries employ new fraudulent methods in increasingly covert and sophisticated ways to maximize profits.
![2024061204.jpg](https://cdn.aisecurius.com/images/6cc/6cc16e27-c900-4a6e-96c1-0e0a813ac17c.jpg)
According to data previously released by Dingxiang Defense Cloud Business Security Intelligence Center, the highest risks occur during shopping events like Double 11 and 618, with November accounting for 39.72% and June for 20.69%.

The marketing cycle for e-commerce platforms spans from half a month to a month before the frenzy day, with marketing investments lasting 1-2 months. This ample time allows black and grey industries to study the processes and rules of various e-commerce platforms in preparation for subsequent fraudulent activities. Popular products and promotions become means for them to profit.


# Coupon farming: Popular products taken away by merchants


During shopping festivals, countless consumers eagerly anticipate, yet many discounted or special-priced items often show "sold out" or "temporarily out of stock". What causes this? It turns out that many popular, limited-sale, or high-value products are snapped up by coupon farmers.

Actions like coupon clipping and grabbing flash sale items not only harm the legitimate interests of consumers but also lead to significant economic losses for consumers, merchants, and e-commerce platforms. Data shows that 70%-80% of e-commerce enterprises' marketing expenses are consumed by coupon farmers, resulting in annual losses amounting to hundreds of billions.

Coupon farmers engage in activities such as grabbing coupons, seizing flash sale items, and registering fake users, causing significant economic losses to consumers, merchants, and e-commerce platforms.

Bulk registration of fake accounts. Registration is a key process for creating an account. Black and grey industries utilize registration tools for automated bulk account creation, registering hundreds to tens of thousands of accounts to facilitate mass purchases and hoarding.

Using group control to manipulate accounts. Black and grey industries utilize group control to operate tens, hundreds, or even thousands of devices from a single computer, performing unified actions such as registration, login, ordering, and purchasing. Group control also offers functions like simulated location, shake-to-refresh, bulk contact list import, and message pushing.

Forging IP locations. An IP address represents a user's network information when online. Black and grey industries use tools for rapid IP dialing, automatically calling dynamic IP addresses across regions, with features like automatic switching, reconnecting after disconnection, automatic browser cookie cache clearance, and virtual network card information, seamlessly switching between domestic and international IPs.

Forging GPS locations. GPS positioning indicates a user's geographical location during online service use. Black and grey industries utilize simulation software and third-party tools to change latitude and longitude, enabling instant "teleportation" to any location.

Forging device attributes. Device attributes like model, serial number, and IMEI are unique identifiers. Black and grey industries use device modification tools to intercept device interfaces at the system level. When applications call these interfaces to retrieve device parameters, they receive attributes forged by the modification tool. Typically, it takes 2-3 minutes for such tools to modify attributes for 1000 devices.

## # How merchants can prevent coupon farming


Business risk fraud constantly evolves with rapid changes in methods. New attack methods adjust existing preventive measures or render them immune. Traditional measures often fail to timely identify and alert new risks, and contradictions among various e-commerce platform security processes sometimes lead to direct impacts on order transactions and store operations.

![2024061205.png](https://cdn.aisecurius.com/images/215/215a050f-5ce6-4a3b-bba0-8981b3a891a8.png)

Identify and alert risk IP addresses. Integrate an IP risk database to match user-associated IPs for risk identification, identifying risks like proxies and rapid dialing IPs, and intercepting malicious IP addresses.

Identify and alert risk devices. Assess whether client device fingerprinting is legitimate, detecting risks like injection, hooking, and simulation software. Quickly recognize risks associated with rooted devices, jailbroken devices, or devices hijacked for injection. Rapidly identify abnormal behaviors such as repeated activations from the same device, abnormal IP clustering within a short time, unusual proportions of old device models or operating systems within the same channel.

Identify and intercept risky accounts. Analyze verification environment information and tokens during validation completion to promptly detect anomalies and risky operations. Monitor abnormal account behaviors. Deploy strategies based on user behavior, focusing on accounts that frequently switch from the same device during order initiation.

Analyze and forecast future behavioral risk changes. Establish and maintain a dynamic operation mechanism for local lists based on registration, login, and activation data, accumulating and maintaining corresponding black and white list data dimensions such as user ID, phone number, and device. After accumulating online data, use risk control and business accumulation data to model registration, login, ordering, and purchasing behaviors, directly incorporating model outputs into risk control strategies.


# Advertising fraud: causing colossal promotional efforts by e-commerce platforms to go down the drain


During online shopping festivals, e-commerce platforms heavily invest in advertising and native advertising to attract more consumers. Unscrupulous individuals exploit various methods such as ad clicking and falsifying promotional data to reap benefits. On one hand, after advertising campaigns, stakeholders only see corresponding click-through rates, activation rates, and registration rates, with very low retention rates. On the other hand, the data includes a large amount of fake traffic with no conversions from real users, and efforts to analyze failed conversions further cannot pinpoint the true reasons. This not only wastes advertising costs, time, and development expenses but also causes substantial asset losses and a plethora of valueless fake users for advertisers.

Analysis by Dingxiang Defense Cloud Business Security Intelligence Center indicates that perpetrators engage in e-commerce advertising fraud mainly through three forms: manual brushing, machine brushing, and wall brushing.

Manual brushing involves using real devices to manipulate key parameters like IMEI, phone numbers, IMSI, MAC addresses, etc., under human operation, to simulate multiple user downloads, activations, registrations, and usage behaviors for cheating traffic.

Machine brushing entails attackers using a large number of phones to repeatedly change their configurations, masquerading as new devices to repetitively "install-activate-delete-install-activate-delete," thereby inflating activation figures. These devices typically exhibit root access, emulator use, VPNs, abnormal charging states, GPS information, and SIM card behaviors, which can be detected and mitigated by combining device information, tracking user behavior, retention rates, and online durations.

Wall brushing involves genuine human and device operations where users on such platforms are real smartphone users. Although many users uninstall the app after completing tasks, there remains a certain conversion rate. Therefore, many advertisers still opt for wall brushing as a promotional method.

Because of actual rewards, users will go to great lengths to earn rewards through wall brushing, fostering fraudulent behaviors to exploit promotion fees. Some users manipulate batches of fake accounts to complete tasks fraudulently, resulting in heavily adulterated promotional trial data and causing significant losses to advertisers.


## How e-commerce platforms can prevent advertising fraud


Based on various advertising fraud methods, Dingxiang Defense Cloud Business Security Intelligence Center recommends that advertisers reinforce terminal protection while leveraging business-side combinations to further safeguard promotional effectiveness.

Ensure client-side security to prevent hijacking by black and grey industries. Dingxiang's adaptive App shielding employs graph neural network technology to analyze and extract code characteristics, automatically selecting suitable methods for obfuscation based on different code blocks to significantly increase the difficulty of reverse engineering, effectively reducing computational performance consumption by 50%. Through encryption and obfuscation engines, it enhances App code security, effectively preventing attacks such as cracking, duplication, and repackaging.

Secure account safety to prevent account theft. Dingxiang's atbCAPTCHA, utilizing AIGC technology, prevents AI-based brute force, automation attacks, and phishing threats, ensuring unauthorized access prevention, account protection from theft, and malicious operations to maintain system stability. It integrates 13 verification methods and multiple prevention strategies, encompassing 4380 risk policies, 112 risk intelligence categories, covering 24 industries and 118 risk types. Its precision in risk control reaches 99.9%, with rapid transformation from risk to intelligence. It supports seamless security verification for users while achieving response times within 60 seconds, further enhancing user login service experience convenience and efficiency.

Identify forged devices to prevent various malicious abnormal behaviors. Dingxiang's Device Fingerprinting integrates multiple device information to generate unified and unique Device Fingerprints for each device. It builds multi-dimensional identification strategy models based on devices, environments, and behaviors to recognize risks like virtual machines, proxy servers, and emulators manipulated maliciously. It analyzes device behavior anomalies such as multiple account logins, frequent IP address changes, and abnormal device attribute changes that do not match user habits, tracking and identifying fraudulent activities to assist enterprises in managing operations under the same ID across all scenarios and channels, enhancing cross-channel risk identification and management capabilities.

Uncover potential fraud threats to prevent various new fraud attacks. Dingxiang's Dinsight aids enterprises in risk assessment, anti-fraud analysis, and real-time monitoring to enhance risk control efficiency and accuracy. Dinsight processes daily risk control strategy with an average speed of less than 100 milliseconds, supporting configurable access and sedimentation of multi-source data. It uses mature metrics, strategies, and deep learning technology to automatically optimize security strategies against known risks, configuring anti-fraud strategies based on risk logs and data mining potential risks. Its standardized data processing, feature derivation, model building, and one-stop modeling service from data processing to final model deployment enhance overall security measures effectively.

Advertising fraud and coupon farming: making the promotion and marketing of e-commerce 618 futile

Enhance business risk control security with our atbCAPTCHA and app security solutions.  Protect against attacks and safeguard user data. No coding required. Get a free trial now!

Why AISecurius?

Compliance

Intelligent

Professional

Products

Use Cases

E-commerce

Financial

Airline

Logistics

Games

Live Stream

Contact Us