Facial recognition technology has become increasingly popular as a security tool, with applications ranging from law enforcement to smartphone security, and even corporate access control. It is being used to keep a close eye on who is accessing which locations or devices.
However, recent research by AI firm Kneron suggests that this technology may not be as foolproof as previously thought.
All it takes to fool facial recognition at airports and border crossings is a printed mask, researchers found https://t.co/42ymWrzYZI #fintech #biometrics pic.twitter.com/n294lfUUmw
— Chris Gledhill | gledhill.eth (@cgledhill) December 13, 2019
Kneron’s researchers were able to fool facial recognition systems at various border crossing checkpoints, banks, and airports by wearing printed masks of other people’s faces. The company conducted tests across three different continents and found that payment tablets run by Chinese companies AliPay and WeChat, as well as facial recognition systems at a border crossing in China and Amsterdam’s Schipol Airport systems in the Netherlands, were all fooled by the printed masks.
While it is concerning that such technology can be fooled by something as simple as a printed mask, it should be noted that the masks used in the tests were not low-cost plastic ones found in costume stores of President Trump or the Queen of England. Nonetheless, the fact that they were able to trick these systems highlights potential flaws in facial recognition technology.
Facial recognition systems have been welcomed with open arms in many airports as they decrease waiting times. The ability to quickly identify people by scanning their faces and comparing them to their passport photos has made border crossing checkpoints more efficient.
This is integral for security reasons; however, it looks like these systems may not be as unfailing as we give them credit for.
It’s worth noting that more sophisticated facial recognition technology, such as light imaging, was not as easily fooled as the systems tested by Kneron. The technology used by Huawei and Apple ID was not fooled when the researchers attempted to trick them.
In conducting the tests, Kneron’s researchers were supervised and had permission from security guards. Additionally, there are usually people who are near the facial recognition machines, so it’s possible that they would notice someone wearing a mask if the occasion were to arise.
Nonetheless, this research suggests that there may be some potential vulnerabilities in facial recognition technology that need to be addressed before it can be relied upon as a foolproof security measure.