Meta Quietly Removed Face-Recognition Code From Its Smart Glasses App

Meta removed face-recognition code from its Ray-Ban Meta smart glasses platform around June 8, 2026, after WIRED discovered the unreleased feature buried in the companion app. Engadget confirmed the removal the same day. The code was never shown to users and was not listed as something the glasses could do — but its presence in the app's underlying code was enough to prompt Meta to take action once the story went public.

What Was Found

Researchers and journalists found face-recognition code sitting dormant inside Meta's AI Glasses companion app. This kind of discovery — searching through app code for hidden features before they launch — has become routine in tech reporting. The Ray-Ban Meta glasses already use AI powered by the device's camera to offer real-time information and suggestions, meaning the basic infrastructure to add face recognition was already in place. Adding facial identification on top of that existing system would be straightforward, technically speaking.

Meta's published privacy page for the AI Glasses makes no mention of face recognition as a current or planned capability. Instead, it describes protections designed to strip out identifying information before AI processing happens. The gap between what Meta's privacy documentation promised and what the hidden code suggested was what made the story newsworthy.

How Meta Responded

Meta removed the code quickly and without explanation. The company never said why the code was there in the first place or what it was meant to do. This kind of silent deletion, rather than being upfront about it, is worth paying attention to. Regulators, especially the FTC and European data-protection authorities, already scrutinize Meta heavily. Companies under that kind of oversight often ask whether unreleased features buried in production code require any kind of public disclosure. In most places today, the answer is no — as long as the feature is never turned on. But once a journalist finds it and the company quietly removes it, that "never turned on" claim becomes harder to believe.

The Ray-Ban Meta glasses do have a physical power button — hold it for three seconds to turn the glasses on or off — and Meta points to this as a privacy safeguard. Whether a button you can manually control is enough protection against a system that could silently identify people's faces is a question that regulators on both sides of the Atlantic have been debating without reaching any clear answer.

Why This Matters More on Glasses Than on Phones

Face recognition on a smartphone requires you to open an app and deliberately point the camera at something. Face recognition built into glasses you wear all day — something your friends probably won't even notice is a camera — is a fundamentally different kind of threat. The core issue is asymmetry: the person wearing the glasses can identify people silently, but the person being identified has almost no way to know it is happening and cannot do anything about it in the moment.

This is not a theoretical worry. In late 2024, Harvard students built a proof-of-concept system called I-XRAY that took a Ray-Ban Meta glasses feed, streamed it online, and used a face-recognition service to instantly identify strangers. They did not need to get inside Meta's code to do it — they just put together publicly available tools. The fact that Meta had face-recognition code built into its own app suggests the company was at least experimenting with doing this natively.

The industry has faced this moment before. When Google Glass launched in 2013, Google banned face-recognition apps from the platform after public outcry. Many people at the time saw that as Meta bowing to social pressure rather than making a permanent decision. Google Glass eventually failed in the consumer market anyway. But smart glasses are back now, with much bigger numbers of users, and the face-recognition question has come back with them — still without any laws or rules that clearly address what these devices can and cannot do.

The Trust Problem

Meta's official privacy page says it removes identifying information before sending video to its AI servers. That is a meaningful protection — if it actually works that way and if someone independent could verify it. The problem is that neither of those things is publicly provable from outside the company. Finding hidden face-recognition code does not prove Meta was secretly using it or planned to turn it on without asking permission. But it does suggest that what Meta says in public about privacy protections and what is sitting in the code behind the scenes do not always line up.

This also matters for businesses buying these glasses for their employees. Companies that buy Ray-Ban Meta glasses for workers who use them for productivity tasks face a real question: can they audit what the camera is doing, the way they can audit a smartphone through standard mobile device management tools. The answer is no, not really. This is a structural gap, and stories like this one tend to make procurement teams realize they never fully thought through it.

What Should Happen Next

Meta has not publicly said where the face-recognition code came from, when it was developed, or whether it might come back in a future version with proper user consent built in. These three questions are the ones that matter most to privacy researchers, regulators, and the developers building on Meta's platform.

The EU AI Act, which started being enforced in 2025, bans real-time facial identification of people in public spaces except in narrow cases. Whether unreleased code sitting in a consumer app triggers any kind of reporting or review requirement is exactly the kind of edge case that regulators are still working out. The United States does not have an equivalent law at the federal level, though state-by-state biometric privacy laws like Illinois BIPA could apply depending on where the data ends up.

There is a more optimistic way to read this. The system did work in one sense: independent reporting found an undisclosed feature, Meta removed it after public pressure, and the episode will probably make regulators pay closer attention and push Meta to tighten its own controls on biometric features. That is a form of accountability, even if it only kicked in after someone had to dig through the code to find it.

The harder problem is one the industry still has not solved since Google Glass. What should govern face-recognition capabilities in wearables before they ship to users — not after a journalist finds the code hiding inside the app. That is the question that still needs an answer.