Meta Sued for Allegedly Harvesting Photos to Train AI Models and Build Secret Face Recognition
nordicinst · x · 2026-09-12
A proposed class action filed in federal court in Chicago alleges Meta illegally used Facebook and Instagram photos to train generative AI models (Emu, Muse Image) and build "NameTag," an unreleased face-recognition system for its smart glasses, violating Illinois and California privacy laws by extracting biometric data without consent.
Key facts:
- WIRED reported in June that NameTag code was secretly embedded in the Meta glasses AI companion app, downloaded over 50 million times, though never enabled.
- The system was designed to turn faces captured by the glasses into biometric signatures matched against faceprints in an on-device database receiving updates from Meta.
- The complaint alleges these faceprints may derive from Facebook/Instagram images, citing employee claims and a Meta patent.
- Meta previously stated it is "not building a central face database."
EU watchdogs are watching closely; the case could push tighter rules and cleaner training data.
More from Safety
- Frontier labs' privacy terms are 'insane': one thumbs-up can void your chat protections — niloofar_mire · 2026-09-12
- VC Rob Leclerc: AI safety researchers should face nuclear-plant-level background checks — robleclerc · 2026-09-12
- Gary Marcus: Hugging Face incident shows harmful infrastructure-attacking AI needs no AGI — GaryMarcus · 2026-09-12
- OpenAI pitches cybersecurity services to keep its AI from hacking the US power grid — ccerrato147 · 2026-09-12
- 76% of viral Anthropic resignation post's reposts came from abroad, fueling astroturf claims — StewartalsopIII · 2026-09-12
- Sanders AI bill proposes up to 20 years in prison for developers of 'advanced AI' — ivan_bezdomny · 2026-09-12