Meta Sued For Using Facebook, Instagram Photos To Train AI For Smart Glasses

Meta is facing yet another lawsuit, this time for allegedly illegally using users' Facebook and Instagram photos for facial recognition.

The case, filed by a set of parents and their children in Illinois and California, alleged that Meta used their Facebook and Instagram photos without consent and fed the data to NameTag, an unreleased facial recognition system for its new AI glasses.

The lawsuit further alleges that the photos were used to train generative AI models, including Emu and Muse Image, without taking their consent.

The 66-page complaint was filed in the US District Court for the Northern District of Illinois last week. At the centre of the lawsuit is NameTag, which was built for Meta's Ray-Ban and Oakley glasses. It also acts as the Meta AI companion app.

NameTag's existence was revealed by a WIRED report in June, when it said that there was an inactive code for the system in the Meta AI app. Analysts at the time determined that the system was designed so that it could turn faces captured by the glasses into biometric signatures or a faceprint. It could then compare it with other faceprints on the wearer's phone.

The lawsuit alleges Meta went well beyond developing the software to merely capture faces. It says that those photos have possibly been taken from Facebook and Instagram images. The complainants cited two Meta employees, who claimed that NameTag could identify people through their public Instagram accounts or connections.

"In other words, NameTag would work by collecting images of faces from user-generated content on Meta's platforms (including Facebook and Instagram), using those images to create biometric identifiers and information associated with those users, storing the biometric identifiers on Meta's servers, and then providing them to NameTag so that they could be used by NameTag to identify people recorded by the company's Meta Glasses," the lawsuit notes.

The complaint further alleges that Meta's generative AI systems, Emu and Muse Image, harvested biometric information of people whose images were used by the company to train them.

In a statement to WIRED, a Meta spokesperson dismissed the lawsuit and called it a 'misrepresentation' of its work. "As for NameTags, nothing has shipped to consumers and no final decision has been made on what to do here, if anything."

The proposed class in the lawsuit includes people in Illinois, California, and across the United States who uploaded their images on Meta or fed data to its gen AI systems, dating back to September 2021. The complaint estimates the national class could exceed millions of people.