WIRED says Meta removed code tied to an unreleased face-recognition system from the Meta AI app, the companion app used with the company’s smart glasses, one day after the publication reported that the system had been quietly embedded in the app.
The finding has not been independently verified by this publication. According to WIRED, its review of the latest app code found that components associated with the internal system, which it said Meta called NameTag, were no longer present in the version released after its report.
Meta has not publicly confirmed that it removed the code because of the report. Andy Stone, Meta’s vice president of communications, told WIRED the feature was exploratory and said, “No final decision has been made.”
The episode has drawn attention because face recognition in wearable cameras carries a different privacy risk than face recognition on a phone or laptop. Smart glasses can record from the wearer’s point of view in ordinary public settings, where bystanders may not know that a device is analyzing faces or preparing data for later identification.
What WIRED Said It Found
WIRED reported that a previous version of the Meta AI app contained several software libraries explicitly named for face recognition. The outlet said those components were connected to NameTag, an unreleased system designed for Meta’s smart glasses.
Because those findings are based on WIRED’s code analysis and have not been independently reproduced here, they should be read as attributed reporting rather than confirmed technical fact.
According to WIRED, the system was not publicly enabled. The outlet reported that the feature appeared designed to turn faces captured by the glasses into biometric templates, often called faceprints, and compare them against face data stored on the user’s device.
WIRED also said it found signs that faces the system did not recognize could be cropped, indexed, and kept locally for later processing. Meta did not answer WIRED’s detailed questions about how long such data would be retained, whether it could be sent to Meta servers, or how users and bystanders would be notified.
Those unanswered questions are central to the privacy concern. A phone-based face-recognition feature generally requires a person to point a device, open an app, or take a visible action. Smart glasses make that boundary less obvious. The camera is worn at eye level, and the people being captured may not understand what is being collected or analyzed.
Meta’s Public Position
Meta’s public response, as described by WIRED, has been that the feature was exploratory and that no final decision had been made about whether to release it. Stone also said the company could not answer some questions about how the system would work because the feature did not exist as a released product.
WIRED reported that Meta chief technology officer Andrew Bosworth criticized the outlet’s earlier coverage, calling it misleading and dishonest. The dispute leaves two separate issues in view: whether Meta had a finished consumer feature, and whether unreleased face-recognition code had already been included in an app installed by a large number of users.
Those are not the same question. A feature can be unreleased and still have meaningful privacy implications if code connected to it is present in consumer software, especially when the software works with cameras and biometric data.
Meta has not said, based on the source material provided, whether NameTag was intended for accessibility, identity assistance, social features, or another use. WIRED said the company did not answer whether the system was being developed specifically for blind or low-vision users. That matters because accessibility uses can be legitimate and valuable, but they still require clear limits around consent, retention, and misuse.
Privacy advocates have warned that face recognition in glasses could be abused to identify strangers in public. That concern is especially sharp in situations involving stalking, harassment, domestic abuse, or political activity, where the ability to identify someone without their awareness could create real-world risk.
What Reportedly Changed In The App
WIRED said its review of the later Meta AI app version found that the main face-recognition components were gone. It reported that the removed material included the recognition process, a user-facing alert that would have appeared when a person was identified, and a folder tied to storage of cropped face images and biometric signatures for people the system did not recognize.
Again, those technical claims come from WIRED’s analysis and have not been independently verified here. Meta did not answer WIRED’s questions about why the code was removed or whether the change had already been planned before the report was published.
WIRED said a few remnants remained, including a debug menu label and a dormant link associated with opening a recognized person’s profile. The outlet characterized those fragments as leftover references to parts of the system that were no longer present.
The practical effect, if WIRED’s account is accurate, is that the visible code trail for NameTag became much smaller after public scrutiny. But removal from one app version does not answer the larger questions: how far the system had progressed, what data it might have handled in testing, and what safeguards Meta would require before any public launch.
Why The Privacy Debate Is Broader Than One App Update
The controversy sits inside a longer argument over biometric privacy. Faceprints are not like passwords. A person can change a compromised password, but they cannot change their face. That makes collection, retention, and sharing rules especially important.
The source article also points to state-level privacy legislation as part of the debate. Kade Crockford, director of the technology for liberty program at the ACLU of Massachusetts, told WIRED that the reported removal did not erase the concern raised by putting such code into consumer software in the first place. Crockford argued that privacy laws need stronger enforcement tools, including the ability for affected people to sue when their rights are violated.
That argument reflects a broader frustration among privacy advocates: companies can test powerful data systems faster than lawmakers can regulate them. When the technology involves wearable cameras and biometric identification, the gap between experimentation and public accountability becomes harder to ignore.
Meta’s smart glasses strategy depends on making cameras and AI assistants feel ordinary in daily life. Face recognition would push that strategy into a more sensitive category. Even if all matching stayed on a user’s device, the system would still raise questions about who is enrolled, who controls the data, and what happens to people who never agreed to participate.
The Unanswered Questions
The reported code removal narrows the immediate technical issue, but it does not settle the policy issue. Based on the source material, Meta has not provided public answers to several key questions:
- Whether NameTag or a similar face-recognition feature will return to the Meta AI app.
- Whether any face profiles or local biometric databases were created during testing.
- How long images or biometric templates of unrecognized people would be stored.
- Whether any such data could be sent to Meta’s servers.
- How users, bystanders, and people captured by smart glasses would be notified or asked for consent.
Until those questions are answered, the reported removal mainly shows that Meta can change the app quickly when scrutiny arrives. It does not show what the company’s final rules for face recognition in smart glasses would be, or whether the system is permanently shelved.
For now, the safest reading is narrow: WIRED says Meta removed code associated with an unreleased face-recognition system from the Meta AI app after WIRED reported on it. Meta says no final decision has been made. The privacy questions around smart glasses, biometric identification, and bystander consent remain unresolved.