Meta strips face-recognition 'NameTag' code, adds Facebook AI mode

Following a WIRED investigation that exposed dormant face-recognition libraries embedded in the Meta AI companion app, Meta's latest release removed nearly all of the 'NameTag' modules capable of generating local faceprints. The discovery had alarmed privacy advocates: even unused, the presence of facial-recognition code in a consumer AI app raised the specter of Meta building toward real-time face identification—a capability with profound surveillance implications.

Meta's decision to strip the code, rather than leave it dormant or activate it, suggests the company calculated the reputational and regulatory risk outweighed any planned feature. Face recognition has been a persistent legal liability for Meta, which previously paid large settlements over biometric-privacy claims. Removing the modules is a defensive move to preempt scrutiny as it expands AI features.

Separately, Meta introduced a new AI mode for Facebook and a creator-focused AI assistant that recommends optimal posting times and summarizes comments—incremental additions aimed at embedding generative AI deeper into its core social products. Meta also restored Threads in Türkiye this week following a Turkish Competition Authority decision, accompanied by new features, ending a regulatory-driven absence.

The juxtaposition is telling: Meta is simultaneously retreating from one privacy-fraught AI capability while pushing generative AI further into Facebook and Instagram. It reflects the tightrope Meta walks—aggressive AI deployment to defend engagement, balanced against acute sensitivity to biometric-privacy backlash. The caveat is that 'removed nearly all' modules leaves ambiguity about what remains, and Meta's history makes privacy advocates wary of dormant capabilities returning. Watch for regulatory follow-up, whether the new Facebook AI mode gains traction, and how Meta's AI companion app evolves under continued scrutiny.