Your Meta AI app was secretly carrying facial recognition code designed to identify strangers through smart glasses. WIRED’s investigation found that Meta had embedded substantial portions of an unreleased system called “NameTag” into the companion app used by tens of millions of Ray-Ban Meta glasses owners. One day after the report went public, Meta quietly scrubbed nearly all traces of the code.
The Hidden System That Almost Launched
NameTag was designed to create personal face databases and alert users when previously seen people appeared again.
The buried code revealed Meta’s plan to turn smart glasses into real-world tagging devices. NameTag would detect faces, crop them, and encode biometric signatures for storage on your phone. When you encountered someone again, the system would notify you that a “person was recognized” — essentially building a personal surveillance network of everyone you’d ever photographed.
Faces the system couldn’t identify would be stored locally “for future processing,” creating an expanding database of strangers tied to your device. Security researchers who analyzed the code described it as technically close to launch-ready, despite being dormant and inaccessible to users.
The Rapid Cleanup Operation
Meta removed face recognition libraries, alert systems, and storage locations for biometric data in the updated app version.
Within 24 hours of WIRED’s exposé, Meta released an app update that gutted NameTag’s infrastructure. Gone were the face recognition libraries, the workflow orchestrating biometric processing, and the storage folder where cropped face images would live. Only debug fragments remain — dormant menu items and profile links pointing to deleted components.
Meta hasn’t explained why the code vanished so quickly or whether the removal was already planned before the investigation went public.
Privacy Questions That Remain Unanswered
Meta’s conflicting statements raise concerns about transparency and data handling in smart glasses development.
Meta’s response has been contradictory at best. Executives dismissed the reporting as “dishonest” while simultaneously calling NameTag “purely exploratory.” Company representatives insist “nothing has shipped to consumers” and emphasize they’re not building a central face database.
The company won’t answer basic questions about data retention, server syncing, or whether existing test data was deleted alongside the code. Privacy advocates warn this episode shows why strong biometric laws with enforcement teeth are essential — companies won’t self-regulate surveillance capabilities that could enable stalking and harassment.
Your smart glasses decisions just got more complicated. Meta’s willingness to ship dormant surveillance infrastructure without disclosure suggests the line between “experimental” and “deployed” is thinner than they’d like you to believe.
