"Meta" has quietly removed "facial recognition" code from its AI companion app for "Ray-Ban" and "Oakley" smart glasses after a "WIRED" investigation revealed the existence of such a feature. The case once again calls into question the company's intentions in the surveillance sphere and its readiness to experiment with "biometric tools" without public notification.
The code, known internally by the name "NameTag", was discovered by the "WIRED" team while analyzing Meta's AI app – a program downloaded over 50 million times and necessary for the smart glasses to function. Only a few days after questions were posed by the media, "NameTag" disappeared from the latest version of the app. "Meta" refuses to specify whether the feature was "experimental", whether it ended up in the code "by mistake", or if it was being prepared for a future "official launch".
What is "NameTag"?
According to data published by "WIRED", "NameTag" consists of three "artificial intelligence models" that sequentially "detect faces", "crop" them, and convert them into biometric "face prints" – numerical signatures describing the unique arrangement of features. Upon activation, the system compares the faces captured by the glasses' cameras against a "local database" on the user's phone, which is set up to receive updates from Meta's servers.
Upon recognizing a person, the app sends a "notification", while "unrecognized faces" are indexed and stored in a folder labeled "pending". Thus, the glasses can effectively turn into a personal catalog of faces – with the potential to be linked to profiles, contacts, and other information sources.
EFF: "a distributed surveillance machine"
The "Electronic Frontier Foundation" (EFF) independently confirmed the presence of the code after conducting its own "static analysis" of the app. "Despite there being a billion reasons not to do this, Meta appears to have built the capability to turn its users into a distributed surveillance machine," commented "Cooper Quintin", a senior staff technologist at the EFF's "Threat Lab".
Another researcher established that upon "manually adding a face" to the app's database, the glasses subsequently "recognize" that face every time it enters the camera's field of view. This suggests that the system is fully functional, not just "residual" or "archival" code.
Meta's reaction: denial and attacks on the media
"Meta" executives reacted sharply to the publications. "Andy Stone", VP of Communications, described the "WIRED" article as "worse than shoddy journalism" and "intellectual dishonesty", calling it "pure clickbait, dictated by a biased position".
CTO "Andrew Bosworth" also criticized the material, describing it as "incredibly misleading". Official spokesperson "Ryan Daniels" adopted a more moderate tone, stating: "We’ve previously said that we are exploring such features, and what you’re seeing is evidence of that exploration. No product has reached consumers, and no final decision has been made on what to do in this case — or whether to do anything at all." He added that Meta is "not building a centralized database of faces".
Privacy and internal memos
Data protection experts cited in the "WIRED" analysis emphasize that the code appears to be "actively developed" rather than archived – specific "API calls" and "data structures" testify to ongoing work and testing. The media also reported on an "internal memo" at Meta, which discusses a possible launch of "NameTag" during a time of "dynamic political environment", when "civil society organizations that would normally attack us would be occupied with other issues".
These revelations sparked a sharp reaction from civil rights organizations such as the "ACLU", "Fight for the Future", and the "Electronic Privacy Information Center". They warn that such a feature could allow "stalkers and abusers" to identify people in public spaces without their knowledge or consent.
History repeats itself: from Facebook to smart glasses
Critics point out that in 2021, "Meta" (then "Facebook") deactivated its "facial recognition system" on the social network and agreed to pay "650 million dollars" to settle a lawsuit related to "biometric privacy violations". Against this backdrop, the reappearance of similar technology in "wearable devices" seems particularly revealing.
According to privacy advocates, the "NameTag" case demonstrates a "repeating pattern": the company formally retreats under pressure from regulators and courts, but continues to experiment with new forms of "biometric surveillance" in other products – this time in devices that users wear constantly on their faces.
A temporary gesture or a real refusal?
The key question remains open: does the removal of the code mean a "real refusal" of development, or does it represent a "temporary reaction" to public discontent. At the moment, Meta refuses to give a categorical answer as to whether the "NameTag" project is finally terminated, frozen, or will be reworked and returned in another form.
The case puts Meta in another dilemma between "innovation" and "privacy". While the company insists that it is not building a "centralized database" and that the feature has not reached end users, the very fact that such a tool appears in the code of a mass-market app reinforces doubts: where does "research" end and the "product" begin – and who controls this boundary.