What Meta’s Hidden NameTag Experiment Really Was
Meta’s hidden NameTag feature was a facial recognition system embedded in the Meta AI companion app for Ray-Ban smart glasses, designed to detect, crop, and encode faces into biometric data so that wearers could be alerted when the system recognized people they had previously captured. WIRED’s investigation, summarized by other outlets, showed that Meta quietly shipped significant facial recognition code to tens of millions of phones through multiple app updates, without explaining that the Meta facial recognition system existed in any form. NameTag’s infrastructure included three AI models: one to detect faces, one to crop them, and one to turn them into biometric signatures. While dormant for regular users, security researchers who reviewed the app said the facial recognition code was close to launch-ready, raising immediate concerns about smart glasses privacy and informed consent.
From Quiet Embedding to Overnight Deletion
According to Gadget Review’s summary of WIRED’s reporting, Meta had already embedded substantial portions of the NameTag facial recognition code into the Meta AI app before it attracted public attention. The code was present across several updates and wired to smart glasses like Ray-Ban Meta and Oakley models, even though the feature remained inaccessible to end users. Once WIRED’s investigation went public, Meta pushed an update within 24 hours that stripped out face recognition libraries, biometric processing workflows, and storage locations for cropped face images. Only traces such as debug menus and broken links appeared to remain. Meta did not clearly explain whether this rapid removal had been planned or was a direct response to scrutiny, leaving an uneasy picture: experimental surveillance features can ship at scale first and be questioned only after outside investigators discover them.
Meta’s Defense: ‘Exploration’ Without a Central Database
Meta describes the NameTag feature as an internal exploration rather than a launched product, arguing that shipping facial recognition code in a mass-market app does not mean it shipped to consumers. A Meta spokesperson told Mashable that “nothing has shipped to consumers and no final decision has been made on what to do here, if anything,” and stressed that the company is “not building a central face database.” In Meta’s telling, NameTag was an on-device experiment for Ray-Ban smart glasses, not an attempt to recreate large-scale biometric systems that have triggered legal trouble in the past. But the existence of working facial recognition code on millions of phones, paired with Meta’s history of paying large fines for biometric privacy violations and Clearview AI’s scraping of Facebook photos, makes that reassurance hard for many users and advocates to accept.
Smart Glasses Privacy: Personal Databases and Silent Surveillance
The NameTag feature, as reconstructed from the facial recognition code, would have turned Ray-Ban smart glasses into wearable tagging tools that build personal face databases over time. The system was designed to detect faces in photos or video, crop them, and store biometric signatures directly on a user’s phone. When the glasses’ camera saw someone again, NameTag could notify the wearer that a “person was recognized,” while faces it could not identify yet would be kept “for future processing.” That workflow blurs the line between convenience and personal surveillance. Even without a central server-side database, storing growing collections of stranger faces on individual devices raises smart glasses privacy issues, including stalking risks, covert tracking in public spaces, and the impact on people who never consented to having their faces turned into data inside someone else’s Meta facial recognition system.
Public Pressure, Policy Gaps, and What Comes Next
Meta’s quick removal of the NameTag facial recognition code highlights how public pressure can still shape powerful companies’ surveillance ambitions. The deletion came only weeks after 70 civil liberties and digital rights groups urged Meta to halt any plans for facial recognition on smart glasses, and it followed mounting criticism of earlier biometric missteps. Yet the company’s sharply worded response to WIRED’s reporting, which executives reportedly called “dishonest” while still labeling NameTag “purely exploratory,” underlines unresolved transparency problems. Key questions remain unanswered: how long any test data was kept, whether any of it was synced to servers, and what rules will govern future experiments. Privacy advocates argue this episode shows that meaningful biometric laws, backed by enforcement, are needed because companies have strong incentives to keep pushing surveillance features until they are caught.
