The universal face can 'crack' many identification systems

Scientists say they use a method called "master face", which uses AI to create a fake face pattern to fool the recognition system. The special thing is that this fake face can copy the identification points of many other people to create a "universal key".

With only 9 faces synthesized by the AI ​​StyleGAN system, it is possible to create "universal faces" for 40% of the world's population. The StyleGAN algorithm is an AI model capable of creating digital images of an unreal person's face.

With AI StyleGAN, the team successfully unlocked more than 20% of the 13,000 faces in the University of Massachusetts dataset in just one test. Even some other tests give higher results.

Today's facial recognition systems all use many predefined sets of signs to identify the user's identity. And this is the loophole for the team to develop a face template that can create many copies of the same as the majority of users.

In addition, the researchers say that their product is capable of fooling some methods of identifying people by being able to connect to deepfake technology.

The scientists' research is a warning about the risks of using facial recognition for social networking accounts. Hackers can apply this method to steal user data.

Samuel Daniel

Update 12 August 2021