Overview
- A WIRED analysis published Thursday found code for an unreleased feature called “NameTag” embedded in the Meta AI companion app that links to Ray‑Ban and Oakley smart glasses.
- Reporters and independent researchers say three AI models for face detection, image cropping, and encoding faces into 2,048‑number biometric “faceprints” were deployed from Meta servers and now reside on users’ phones.
- Meta says no face‑recognition feature has been released to consumers and that no final rollout decision has been made, and the company also states it is not building a central face database.
- Outside researchers reproduced parts of WIRED’s findings and tested the pipeline, with one researcher triggering a “Person recognized” notification using a test faceprint, which shows the system appears close to functional.
- The discovery revives past controversies — Meta shut its large photo‑tagging system in 2021 and paid large settlements over biometric data — and raises unresolved questions about consent, who would be enrolled, where faceprints would come from, and how regulators will respond.