According to the Vietnam News Agency, one of the common tricks today is sending threatening emails, demanding payment of fake fees such as traffic fines, unpaid service fees or access to illegal documents. These emails often lead to fake websites, tricking users into providing login information, passwords or banking data. While previously phishing emails were often easily detected due to spelling mistakes or poor quality images, now with the help of AI, they are written in a polished, personalized and convincing way.
According to Christophe Axen, Deputy Director of the Regional Computer Crime Unit of the Belgian Federal Judicial Police in Liège, generative AI is now capable of simulating an entire conversation, catching users off guard. In the same vein, Professor Benoît Donnet (University of Liège) warns that AI can be trained to maintain a convincing tone for a long time, thereby gaining the trust of victims before tricking them into providing information or opening malicious files.
Particularly dangerous is AI’s ability to personalize email content. Cybercriminals can use names, addresses, job titles, and social media data to create fake information. In many cases, fake messages are sent from acquaintances or superiors, making victims even more vulnerable.
AI Fraud: Anyone can be a victim. (Photo: Internet) |
Not just limited to individuals, cybercriminals now operate in a professional organizational model, with a board of directors, technical department, accounting department, and even a separate dining room. They target large targets such as hospitals, businesses, and government agencies. When supported by AI, attacks become more unpredictable - from simulating emails of leaders to spreading documents containing malicious code.
Another form of fraud that is also on the rise is Deepfake - a technology that fakes images and voices using AI. According to VnEconomy magazine, cyber security expert Ngo Minh Hieu of the National Cyber Security Monitoring Center (NCSC) warned that people have been using Deepfake to steal videos , images, cut and paste, and impersonate relatives and friends to defraud money transfers. Deepfake videos can be identified by stiff facial expressions, audio that does not match the image, unusual lighting, or situations such as requesting money transfers and then suddenly disconnecting.
Mr. Ngo Minh Hieu advised people to be cautious with any strange messages on social networks. If there is a request to transfer money, call or Facetime directly and ask private questions that only acquaintances know. Current AI cannot perfectly simulate a real-time conversation with natural emotions like humans.
Security expert Vu Ngoc Son also commented that AI fraud will be the main trend in cybersecurity in the near future. Bad actors will continue to exploit new technologies to deceive users in increasingly sophisticated and unpredictable ways. To prevent this, users should absolutely not provide personal information, log into strange websites or install software of unknown origin. In addition, state agencies need to step up propaganda and guidance on identifying high-tech fraud. Network operators and service providers also need to regularly update technological solutions to promptly detect and prevent fraudulent behavior, ensuring safety for users.
Source: https://thoidai.com.vn/lua-dao-thoi-ai-ai-cung-co-the-la-nan-nhan-213974.html
Comment (0)