What the NameTag Leak Reveals About Facial Recognition Smart Glasses
Meta’s NameTag incident refers to the discovery, exposure, and rapid removal of dormant facial recognition code inside the Meta AI companion app used with Ray-Ban smart glasses, highlighting how undisclosed biometric capabilities can be shipped at scale while officially remaining “experimental” and raising new questions about everyday surveillance through consumer wearables. Researchers found unused code that could turn Ray-Ban smart glasses into facial recognition smart glasses, years after Meta announced the shutdown of Facebook’s facial recognition system in 2021 and deletion of more than a billion stored faceprints. Embedded in multiple app updates since January 2026, NameTag was inactive but technically close to launch, according to security researchers. The app, installed more than 50 million times, quietly carried this infrastructure. Facial recognition has long been controversial, and Meta’s history of regulatory scrutiny and large biometric settlements magnifies concern that hidden device features may outpace public privacy commitments.
Inside NameTag: How Meta NameTag Surveillance Was Supposed to Work
The buried NameTag system was designed to turn smart glasses into a personal tagging engine for the physical world. Code analysis shows it would capture faces through the glasses’ camera, crop them, and encode biometric signatures before storing them on the connected phone. When the wearer saw someone again, NameTag could trigger a notification that a “person was recognized,” enabling a private, device-level face log. Faces that could not be matched would be stored locally “for future processing,” meaning an ever-growing archive of strangers’ images could accumulate on a user’s device. Researchers also found references to three AI models for face detection and processing, plus traces of a user-facing feature to help people remember individuals they had met. One test notification even relied on a template using philosopher Michel Foucault’s face, underscoring how close this unreleased Meta NameTag surveillance system was to being usable in the real world.
From Exposure to Deletion: Meta’s 24-Hour Cleanup Operation
WIRED’s reporting triggered an unusually fast response. Within 24 hours of the exposé going public, Meta pushed an update that removed most of NameTag’s infrastructure from the Meta AI app. Security researchers noted that the company stripped out face recognition libraries, the workflow that coordinated biometric processing, and the storage folders meant to hold cropped face images. Only scattered debug fragments and broken menu items remain, pointing to components that no longer exist. Officially, Meta insists “nothing has shipped to consumers” and describes NameTag as “purely exploratory” internal experimentation. Executives even criticized the reporting as “dishonest” while simultaneously confirming that the code was not intended for immediate release. The speed and completeness of the cleanup, however, suggests the system was mature enough that Meta felt compelled to erase it quickly once its presence in the Ray-Ban smart glasses ecosystem became public knowledge.
Ray-Ban Privacy Concerns and the Gap in Meta’s Public Promises
The NameTag episode exposes a gap between Meta’s public privacy commitments and what its product code reveals. In 2021, the company announced it was shutting down Facebook’s facial recognition system and deleting more than a billion faceprints after years of criticism and biometric lawsuits. Yet years later, dormant facial recognition logic appeared in an app required for Ray-Ban Meta smart glasses and installed tens of millions of times. Meta says it is not building a centralized facial recognition database and would move carefully and transparently if it ever launched such a feature. But the company has not answered key questions about how test data was handled, whether any biometric information synced beyond local devices, or how long any experimental data might have been retained. For many users, Ray-Ban privacy concerns now extend beyond camera etiquette to whether hidden device features can quietly redefine what their glasses are capable of doing.
Hidden Device Features and the Future of Consumer Surveillance
NameTag’s rise and disappearance shows how easily dormant surveillance capabilities can be shipped inside everyday products. If large portions of a face recognition workflow can live for months inside a widely used companion app without disclosure, it raises a broader question: what other hidden device features might be waiting inside phones, wearables, and smart glasses, ready to be activated with a simple update? Privacy advocates argue that this episode proves self-regulation is not enough when companies can deploy near-launch-ready biometric systems under the label of internal testing. The line between “experimental” and “deployed” becomes thin once code is sitting on millions of devices. For consumers, decisions about smart glasses adoption now hinge not only on visible features but on trust that latent capabilities—especially those involving faces and identities—will not quietly expand the reach of everyday surveillance.
