Researchers with an artificial intelligence firm mentioned they had been capable to fool facial-recognition computer software at an airport and mobile-payment kiosks working with a printed mask, highlighting safety vulnerabilities.
two min study
This story initially appeared on Business enterprise Insider
Facial recognition is getting extensively embraced as a safety tool — law enforcement and corporations alike are rolling it out to maintain tabs on who’s accessing airports, shops, and smartphones.
As it turns out, the technologies is fallible. Researchers with the artificial-intelligence firm Kneron announced that they had been capable to fool some facial-recognition systems working with a printed mask depicting a various person’s face.
The researchers, who tested systems across 3 continents, mentioned they fooled payment tablets run by the Chinese businesses Alipay and WeChat, as effectively as a technique at a border checkpoint in China. In Amsterdam, a printed mask fooled facial recognition at a passport-handle gate at Schiphol Airport, they mentioned.
The researchers mentioned their findings recommended that a individual who prints a lifelike mask resembling somebody else could bypass safety checkpoints to fly or shop on their behalf.
“Technologies providers need to be held accountable if they do not safeguard customers to the highest requirements,” Kneron CEO Albert Liu mentioned in a statement. “There are so lots of businesses involved that it highlights an business-wide concern with substandard facial recognition tech.”
Connected: What Entrepreneurs Need to have to Know About Facial Recognition Technologies
Some facial-recognition computer software proved impervious to the printed-mask test, even so. The researchers mentioned Apple’s Face ID and Huawei’s technique passed each use a lot more sophisticated technologies identified as structured light imaging. Kneron mentioned its personal facial recognition computer software also passes the test.
Researchers mentioned that tests at safety checkpoints had been carried out with the permission of safety guards supervising them — suggesting that as lengthy as humans are present to notice the mask, facial-recognition checkpoints are not completely unsecured.
In the month soon after its mask study went viral, Kneron announced that it raised $73 million from investors like Alibaba, Qualcomm, and Horizons Ventures.
“We are excited to continue our journey with partners like Horizons Ventures who share our passion and dedication towards our mission to allow AI on any device [and] democratize AI,” Liu told Business enterprise Insider soon after the fundraising was announced.
Here’s the pitch deck Kneron made use of to raise $73 million.
Connected: 17 Remarkable — and Surprising — Uses of Facial Recognition Technologies