Meta Ray-Ban Glasses Facial Recognition Quietly Removed After Outcry

Functional facial recognition infrastructure was embedded in the Meta AI companion app for Ray-Ban Meta smart glasses and distributed to millions of users before Meta quietly stripped it out less than 48 hours after public exposure, EFF reported two days ago. The Ray-Ban Meta facial recognition code was present in the app and verified as working by researchers. It was not yet accessible to ordinary users as a consumer-facing feature. Those are two different things.

Meta has since refused to answer Wired's questions about whether it plans to bring the feature back, or what happened to any data that may have been collected during internal testing, per EFF. The company has not addressed those questions publicly.

This was not a surprise experiment. The New York Times reported on Meta's plans to add facial recognition to its glasses in early February, and an internal planning memo had been circulating before that, according to the Journal of High Technology Law.

What the Meta Ray-Ban glasses facial recognition code actually did

The feature was called "Name Tag" internally. It would have let glasses wearers identify strangers in real time and retrieve information about them through Meta AI, the Journal of High Technology Law reported last month. The companion app stored face data as vectors of 2,048 numbers representing a person's facial geometry, EFF explained last week. When activated, it would have converted every new face entering the glasses' field of view into that numerical signature and compared it against enrolled profiles.

The code was not theoretical. A researcher manually added a face to the app's database by connecting a phone to a computer in debug mode and issuing commands; the glasses subsequently flagged that person when they came into view, per EFF. That was a working demonstration of the full pipeline. EFF's Threat Lab independently confirmed the code's presence through static analysis, and Wired's reporting confirmed the code was active in the app, though not yet accessible as a standard consumer feature, EFF noted.

EFF argues that what separated these glasses from a live identification tool was a configuration setting, not an unsolved engineering problem. The June 5 app update removed the recognition engine, the "Person recognized" alert logic, and the associated machine learning models and biometric databases, all in a single unannounced release, per EFF. Meta said nothing about it publicly.

Why Ray-Ban Meta privacy concerns escalated before the code was even verified

A coalition of 75 organizations led by the ACLU wrote to Mark Zuckerberg two months ago warning that glasses equipped with this technology could identify strangers at protests, medical clinics, and workplaces, then cross-reference those identities against databases holding information about a person's job, health, relationships, and daily habits, per the ACLU. "The principle here is quite simple: Your glasses should not know my name," said Cody Venzke, senior staff attorney at the ACLU.

Anyone within the glasses' field of view would have no practical way to know they were being scanned, much less refuse it. That puts the feature in direct conflict with Illinois' Biometric Information Privacy Act, which requires affirmative written consent before biometric data can be collected from an individual, according to the Journal of High Technology Law. Most surveillance requires some affirmative act: a search, a post, a login. Glasses that passively scan every face in their vicinity remove that friction entirely. Bystanders never get a chance to say no.

Survivor advocates flagged a more specific danger: passive real-time identification gives stalkers and abusers a low-effort tool for locating people who have deliberately tried not to be found. "Stalking is an extremely common tactic used by perpetrators, and an instant identification feature could place survivors in harm's way by enabling abusers to locate and track them," a spokesperson for Refuge and Women's Aid told researchers, per the Journal of High Technology Law. No incidents directly involving these glasses have been reported, but the underlying harm pattern is well-documented in adjacent contexts.

The coalition, which the ACLU described as calling the technology "a red line society must not cross," demanded Meta publicly disavow plans to equip its glasses with facial recognition, per the ACLU. EPIC and EFF separately urged privacy regulators and state attorneys general to open investigations, and sent letters to the FTC citing safety and privacy concerns, according to the Journal of High Technology Law.

Meta has done this before, and the rollback didn't come with answers

EFF argues that Meta's prior retreats from facial recognition came only after legal and financial pressure, not internal course corrections, per EFF. In November 2021, the company shut down the Facebook feature that scanned faces in every photo uploaded to the platform and pledged to delete more than a billion face templates. The shutdown followed years of legal exposure, not a change of heart.

The settlements tell that story. Facebook settled a sweeping FTC privacy investigation for $5 billion in 2019. In early 2021, the company agreed to a $650 million class-action settlement under Illinois' BIPA. In July 2024, Meta paid $1.4 billion to settle Texas claims arising from that same discontinued face-scanning system. Taken together, those actions represent more than $7 billion in privacy-related liability, according to EFF and the ACLU coalition letter.

The internal planning memo makes the June 5 rollback harder to read as a principled decision. According to documents cited by EFF and referenced in the ACLU's letter, both likely drawing from the same upstream document, Meta's internal planning described wanting to launch facial recognition "during a dynamic political environment where many civil society groups that we would expect to attack us would have their resources focused on other concerns." The coalition called that language "frankly shameful." EFF characterized the removal as a direct response to Wired's reporting and public scrutiny, not an internal decision, per EFF. Each prior reversal was reactive and unannounced, followed by silence rather than explanation. The June 5 update fits that mold exactly.

What's still unresolved

Meta has refused to answer Wired's questions about whether the Meta NameTag facial recognition system will return, or what happened to any data that may have been collected during internal testing, per EFF. Current reporting cannot establish how broadly the code was distributed, whether any real-world biometric data was captured, or whether the rollback applied globally.

For anyone who owns or is considering Ray-Ban Meta glasses, three things can be stated with confidence based on current reporting: the facial recognition code has been removed from the current app version; Meta has made no public commitment not to relaunch it; and no independent body has verified whether any data collected during testing was actually deleted.

Meaningful accountability would require a public commitment not to relaunch the feature, external verification of any data deletion, a regulatory investigation with binding findings, and transparency about the scope of internal testing. Meta has offered none of those. The code is gone from the app. The questions it raised are not.