Testing Facial Recognition Systems with High-Quality Replica Masks: WeChat, Alipay, and Airport Security All Fooled

2019-12-14

Facial recognition technology has been widely regarded as a security tool and is widely used by governments and companies. However, it has been proven that this technology is not reliable. Researchers from artificial intelligence company Kneron said on Thursday that they were able to deceive certain facial recognition systems using a highly realistic mask. They tested the systems on three continents and claimed that they successfully deceived the payment systems operated by Alipay and WeChat, as well as the system at Chinese border checkpoints. These systems are similar to those seen at airports, using facial recognition instead of passwords or fingerprints to verify users' identities. In theory, such masks could allow fraudsters to use someone else's facial information and bank accounts to make purchases.


Even more shocking, the Kneron team successfully deceived the security system at Schiphol Airport, the largest airport in the Netherlands, using only a photo on a mobile phone screen at the self-check-in terminal.


While security agencies have adopted facial recognition as a means to save costs and improve efficiency, the vulnerabilities of these systems have also raised concerns about terrorism. The effectiveness of this deception lies in the fact that the masks already contain all the facial features required by facial recognition systems. However, Kneron admits that such fraudulent behavior is unlikely to spread widely, as the masks used in the experiment were obtained from a military special mask manufacturer.


The researchers also pointed out that Apple's Face ID and Huawei's system were not deceived, as both use complex technology called structured light imaging. The tests at security checkpoints were conducted under the supervision of security personnel, indicating that as long as someone is present to pay attention to the masks, facial recognition will not be completely insecure.Kneron gained insights into the limitations of this technology through experiments and developed its own facial recognition technology. The company, supported by well-known investors including Qualcomm and Sequoia Capital, is creating a system called "Edge AI," which is an artificial intelligence tool that identifies individuals entirely on the device, rather than through cloud-based services. Albert Liu, CEO of Kneron, said in a statement, "If technology providers do not guarantee that users meet the highest standards, they should be held accountable."


The reliability of facial recognition and artificial intelligence has been called into question. Computer scientists recently deceived facial sensors using images from Facebook in a similar experiment to Kneron's. Additionally, artificial intelligence has produced tools that can easily reproduce other people's fingerprints, further proving that the biometric recognition tools we increasingly rely on are much less secure than many people think.

RECENT AI NEWS

RECENT AI TOOLS