Recently, the Telecommunications Cybercrime Investigation Bureau of Baotou City Public Security Bureau announced a case of telecommunications fraud using intelligent AI technology. Mr. Guo, the legal representative of a technology company in Fuzhou, was defrauded of 4.3 million yuan in 10 minutes.
A man was defrauded of 4.3 million yuan through a "friend" video
At noon on April 20, Mr. Guo’s friend suddenly contacted him via WeChat video. His friend was bidding in another place and needed a deposit of 4.3 million, and required a transfer from a public account to a public account. He wanted to borrow Mr. Guo’s company account to transfer the money. Based on his trust in his friend and having verified his identity via video chat, Mr. Guo transferred 4.3 million to his friend's bank card in two installments without verifying whether the money had arrived. Mr. Guo found out he had been deceived when he called his friend's number. The scammer used intelligent AI face-changing and onomatopoeia technology to pretend to be a friend and defrauded him.
"He never mentioned borrowing money to me from the beginning to the end. He just said that he would call me the money first and then ask me to transfer it to his friend's account. Moreover, he made a video call to me at that time. I also said in the video I identified the face and voice, so I relaxed my guard," Mr. Guo said.
Fortunately, after receiving the alarm, the police banks in Fuzhou and Baotou quickly activated the payment stop mechanism and successfully stopped payment and intercepted 3.3684 million yuan. However, 931,600 yuan was still transferred, and efforts are currently being made to recover it.
Someone was defrauded of 4.3 million yuan in 10 minutes. "AI face-changing" scams occur frequently, so be careful!
How to easily identify “AI fake faces”?
In recent years, "AI face-changing" scams have continued. Criminals use "AI voice-changing" and "AI face-changing" technologies to disguise themselves as specific people and conduct video calls with others in real time. The changed facial expression is natural and fake, and the person can pretend to be someone else to contact the victim, gain the victim's trust and then commit fraud.
In the era of artificial intelligence, what we see with our eyes may not be true. The academic term for "AI face-changing" technology is called deepfake, which is deep learning + fake. It is a combination of machine deep learning technology and face impersonation.
However, AI technology is not the original sin, but the people who use technical means to conduct illegal operations. "AI face-changing" detection technology has a variety of methods. For the public, it can be easily identified based on the bugs of "AI fake face".
First, pay attention to the texture characteristics of the “AI fake face”. For example, the details of the eye or tooth outlines of the forged video characters are easily inconsistent; the colors of the pupils of the two eyes are different or the details reflected in the center of the pupils are different; or many forged videos are fake because the video resolution is lower than the original video resolution. The edges of the teeth in the video are too neat, etc.
Secondly, pay attention to the biological characteristics of the “AI fake face”. The "AI fake face" may not conform to the physiological characteristics of normal people. For example, a healthy adult usually blinks once every 2-10 seconds, and each blink takes 0.1-0.4 seconds. In the fake video, the blink frequency of the person is may not comply with the above rules.
Again, pay attention to the inherent characteristics of the mouth of the “AI fake face”. For example, because the movements of the human mouth are the most frequent and rapid, AI software cannot render its continuous movements realistically and accurately. Based on this, it can be used as one of the factors to judge the authenticity of the video.
In addition, the forged video will cause a certain amount of video jitter, resulting in inconsistencies between frames in the video.
As ordinary people, everyone must strengthen the maintenance of their own information security, and do not easily provide personal biometric information such as faces, fingerprints, etc. to others in work and life; when using electronic devices, do not overly disclose or share animal information. Pictures, videos, etc.; secondly, do not click on unfamiliar links, do not download unfamiliar software, and do not add unfamiliar friends casually to prevent viruses on mobile phones and computers, and personal social accounts being hacked; finally, when it comes to money transactions, no matter who the other party is, It is necessary to conduct multiple verifications through phone calls and other methods as soon as possible. Such verification is often the easiest for scammers to reveal their flaws.
