What NameTag Is and How It Reached 50 Million Phones
Meta’s NameTag facial recognition code refers to a set of AI models and libraries inside the Meta AI app that could convert faces into local “faceprints” on users’ phones and link them to in‑app identity features, raising major concerns about facial recognition privacy and mobile app surveillance. According to WIRED’s code review, core NameTag components were shipped through Meta AI app updates starting in January 2026, not kept only in internal builds. By the time the issue surfaced, the companion app had been installed more than 50 million times, meaning those phones quietly received dormant facial recognition modules without a public feature launch or clear user opt‑in. These phones support Meta’s smart‑glasses ecosystem, including Ray‑Ban and Oakley frames, giving Meta a ready path to device‑side identification even before regulators or the public had a chance to weigh in.
How the Dormant Code Worked and Why It Alarmed Researchers
Reverse‑engineering by WIRED and outside researchers found three AI models and interface traces designed to power NameTag: face detection, face cropping, and encoding into faceprints. The code suggested a pipeline where captured faces could be processed locally and matched to people profiles, then used to notify wearers when someone nearby was recognized. While Meta described the work as exploratory, the presence of a debug menu and dormant links to person profiles showed a system that was, in one researcher’s words, “nearly ready to go.” Privacy advocates warned that once such a tool exists on mass‑market devices, flipping it on could normalize real‑time identification in public spaces. Even if NameTag never fully launched, the ability to store and update faceprints on users’ phones raised urgent questions about Meta AI app security and informed consent.
NameTag Code Removal and the June 2026 App Update
After WIRED revealed the dormant NameTag libraries, Meta pushed a new Meta AI app release in June 2026 that stripped out the main facial recognition components. Glass Almanac reports that the latest version “no longer includes the NameTag modules that could have created local ‘faceprints.’” The update removed key recognition libraries and most traces of the debug pathways, signaling a deliberate rollback rather than a routine refactor. Yet Meta spokespeople declined to answer specific questions about whether any images or faceprints had been retained, how long any local data persisted, or if NameTag might return in another form. For many users, the NameTag code removal brought partial relief but not resolution, since the company has not provided a public audit of what was shipped, tested, or stored while the modules sat dormant on tens of millions of phones.
Privacy, Regulatory Fallout, and What Users Should Do Now
The NameTag episode underscores how quickly AR‑driven mobile app surveillance can scale when experimental code ships ahead of public debate. Privacy groups warned that hidden facial recognition capabilities risk enabling stalkers and expanding biometric ID without consent, while civil liberties advocates pointed to Meta’s history of large biometric settlements as evidence that stronger legal safeguards are needed. State legislatures were already drafting broader consumer privacy bills, and this discovery has sharpened calls for stricter oversight of facial recognition privacy in everyday apps. For users, the practical steps are straightforward: update the Meta AI app so NameTag components are removed, review camera and microphone permissions, and decide whether pairing smart glasses with the app still aligns with personal risk tolerance. Longer term, expect regulators and courts to press for transparent data retention policies and explicit opt‑in for any future face‑recognition features.
