Meta Sued Over AI and Face Recognition Training Data
Original: Meta Sued Over Training Data for Its AI and Face-Recognition Systems
Why This Matters
Biometric privacy litigation against a top-5 tech company could reshape how platforms use social media data for AI training.
A proposed class action filed in federal court in Chicago alleges Meta illegally harvested Facebook and Instagram photos to train AI image models Emu and Muse Image, and to build NameTag, an unreleased face-recognition feature for its smart glasses, violating Illinois and California privacy laws.
Parents and children from Illinois and California filed a federal lawsuit last week alleging Meta extracted biometric data from their social media photos without notice or consent. The suit targets two distinct systems: the generative AI models Emu and Muse Image, and NameTag, an unreleased face-recognition feature discovered embedded in Meta's glasses AI companion app — which has been downloaded more than 50 million times.
WIRED first reported on NameTag in June, finding code secretly embedded in the app that converts faces captured by smart glasses into biometric signatures and compares them against faceprints stored locally on a user's phone — with that database configured to receive updates from Meta. The plaintiffs allege those faceprints may be derived from Facebook and Instagram profile photos, citing internal employee claims and a company patent describing face matching against Meta-held images.
On the generative AI side, Meta's chief product officer Chris Cox has publicly described Facebook and Instagram as a 'data advantage' for training Emu. Muse Image, released this summer, was pulled back within days after enabling users to generate images from other people's public Instagram accounts. Meta called the lawsuit 'without merit,' adding that NameTag has not shipped and no final decision has been made. 'We are not building a universal face database,' a spokesperson said.