Meta Silently Added Face-Recognition Code for Its Smart Glasses to Millions of Phones
WIRED found Meta had already shipped face-recognition code for its smart glasses to millions of phones, though the feature is not yet enabled.
Intelligence analysis by GPT-5.4 Mini
The live Meta AI app contains an unreleased system, internally called NameTag, that can identify people seen by Meta smart glasses using biometric data stored on users’ phones. WIRED says the code was added while Meta publicly said it was only still considering face recognition.
Meta appears to have put the parts for a face-matching tool into its smart glasses app, like building a lock before telling anyone what door it opens. If turned on, the glasses could recognize faces and tell the wearer who someone is.
Analysis
What WIRED found
WIRED says its code review of Meta’s live AI companion app uncovered an unreleased face-recognition feature for Meta smart glasses, internally called NameTag. The feature is embedded in an app downloaded by more than 50 million people and used for Ray-Ban and Oakley smart glasses features.
How it is described to work
According to the article, the system would take faces seen by the glasses’ camera and turn them into biometric signatures, often called faceprints. Those faceprints would be checked against data stored on the user’s phone, and the database is described as currently configured to receive updates from Meta. Recognized people would trigger a notification, while unrecognized faces would be cropped, indexed, and placed in a folder marked pending.
Why the discovery matters
The article says Meta publicly described face recognition as something it was still thinking through, including a statement in April that any rollout would require a thoughtful approach. WIRED says it found core parts of the feature in software distributed as early as January. The piece also notes that Meta had said it shut down its earlier face-recognition system in 2021 after years of controversy and legal pressure.
Evidence and reaction
WIRED says outside researchers reproduced key parts of its analysis. One researcher, Cooper Quintin of the EFF, said the feature was not yet exposed to consumers but seemed nearly ready. Another researcher, Buchodi, reportedly added a known faceprint and triggered a "Person recognized" notification, suggesting the main recognition pipeline is already present.
The policy and safety question
More than 70 advocacy groups had already urged Meta to scrap the feature, warning it could help stalkers and abusers identify strangers in public. Meta’s spokesperson said the company was exploring the idea, that nothing had shipped to consumers, and that no final decision had been made. The article also notes Meta says it is not building a central face database.
Key points
- WIRED says Meta’s smart glasses app already contains unreleased face-recognition code called NameTag.
- The feature would compare faces seen by the glasses against faceprints stored on the user’s phone.
- Recognized faces would trigger alerts, while other faces would be cropped, indexed, and saved as pending.
- Meta publicly said it was still thinking through face recognition and that nothing had shipped to consumers.
- Security researchers reproduced parts of the analysis and said the feature appears close to working.
If Meta limits the feature tightly and makes it opt-in with clear controls, it could help people who struggle to remember names or faces, including blind users. The article says Meta has framed the technology as something it is still exploring, so a cautious rollout could still be shaped by stronger privacy protections.
The article’s main concern is that a face-recognition feature inside consumer glasses could become a silent identification tool in public. Privacy advocates warn that this could help stalkers, abusers, or other users identify strangers without their knowledge, normalizing a technology Meta previously retreated from.
