PimEyes, a public search engine that uses facial recognition to match people's photos online, has now been banned from minors due to concerns about children's personal safety, according to The New York Times. At least in theory. PimEyes' new detection system, which uses age-detection AI to identify whether a face is that of a child, is still in development. After testing it, the New York Times found that it was still difficult to identify whether the person being photographed was a child at certain angles, and the AI could not accurately detect whether a teenager was an adult.
PimEyes CEO Giorgi Gobronidze said the company has been planning to implement such protective measures since 2021. But after New York Times writer Kashmir Hill published an article last week about the threat AI poses to children's personal safety, the feature was finally fully deployed.
Hill wrote in the article that the service has banned more than 200 accounts for making inappropriate searches for children. One parent told Hill that she even discovered photos of her children that she had never seen using PimEyes before. To find out where the images came from, the mother had to pay a $29.99/month subscription.
PimEyes is just one of the facial recognition engines in the world currently under scrutiny for violating privacy. In January 2020, Hill's New York Times investigation revealed how hundreds of law enforcement organizations used a facial recognition engine called Clearview AI with little oversight.
Daly Barnett, a technology expert at the Electronic Frontier Foundation, told The Intercept last year: "This is just another example of a major overall problem in technology, whether it has surveillance capabilities or not." He also criticized PimEyes' lack of child protection measures at the time: "[PimEyes] did not build privacy protections from the beginning, and users have to choose not to have their privacy compromised."