What Meta’s Hidden ‘NameTag’ System Was Designed To Do
Meta facial recognition in the Meta AI app refers to dormant code, internally called NameTag, that could turn Ray-Ban smart glasses into a tool for biometric data tracking, creating faceprints on a user’s phone and silently matching people in view without their clear knowledge or consent. Investigators found that NameTag was built to detect faces seen through Meta’s smart glasses, crop those images, and convert them into unique biometric signatures stored locally. When the wearer encountered a person again, the app could flag that a face had been recognized, effectively creating a personal, always-on tagging system. Security researchers say the facial recognition code path included models for face detection and matching, plus a local database architecture meant to live on the phone while still receiving updates from Meta’s servers. Even dormant, this architecture made smart glasses surveillance technically possible.
From Silent Shipments to a 24-Hour Cleanup
According to reporting confirmed by independent researchers, Meta embedded facial recognition code into its Meta AI companion app over multiple updates starting as early as January. That app underpins Ray-Ban Meta glasses and has been downloaded more than 50 million times, meaning NameTag’s components quietly reached millions of phones even though the feature remained off. WIRED’s investigation showed that the code path was close to launch-ready, with workflows for detecting faces, creating faceprints, and sending recognition alerts. After the investigation went public, Meta pushed an app update within about 24 hours that removed face recognition libraries, biometric storage folders, and alert systems, leaving behind only scattered debug references. This rapid reversal undercuts the idea that the project was a distant experiment and instead suggests a feature that had progressed well beyond a rough prototype, raising new Ray-Ban privacy concerns.
Meta’s Explanation—and Its Facial Recognition History
Meta insists that NameTag never launched for consumers and that the remaining code was evidence of internal experimentation rather than a planned feature. A company spokesperson wrote in an email that no final decision had been made and that Meta is not building a central face database. Yet the Electronic Frontier Foundation’s Threat Lab describes the design as capable of turning customers into a “distributed surveillance machine” by normalizing smart glasses surveillance without bystanders’ consent. Meta’s history makes critics wary. The company once ran a massive facial recognition system on Facebook photos to power Tag Suggestions, then shut it down and deleted more than 1 billion faceprints after legal and regulatory backlash. Meta also agreed to pay USD 650 million (approx. RM2,990 million) to settle a lawsuit over biometric privacy, followed by a later USD 1.4 billion (approx. RM6,440 million) settlement in Texas.
Why Dormant Code Still Matters for Privacy
Even though NameTag remained switched off, its presence inside a widely used companion app shows how far hidden surveillance features can spread before anyone notices. Dormant facial recognition code still defines what Ray-Ban smart glasses are technically capable of doing, especially when that code can be activated with a configuration change rather than a full new build. Privacy advocates worry that this normalizes biometric data tracking in everyday life, where people in public spaces cannot meaningfully consent to being scanned. The episode underscores a deep gap between public promises and buried implementations: Meta publicly promotes privacy protections for its smart glasses while shipping an app that quietly contains a near-complete facial recognition system. For users, the lesson is that transparency cannot rely on marketing pages or settings screens alone; it requires independent scrutiny of the software running on their devices.
