Meta's Smart Glasses Can ID Strangers in Seconds. 75 Groups Say Kill It Now.

A security researcher walked into the R.S.A.C. conference in twenty twenty-six wearing a pair of Meta Ray-Ban smart glasses. Within seconds, those glasses — paired with a commercial facial recognition system — identified strangers by name and pulled up their social media profiles. No one being scanned knew it was happening.

More than seventy-five civil liberties

Now more than seventy-five civil liberties organizations have signed an open letter demanding Meta kill its planned facial recognition feature for those glasses entirely. Not redesign it. Not add an opt-out button. Kill it. The coalition argues this technology — quote — "cannot be resolved through product design changes, opt-out mechanisms, or incremental safeguards." And whether you investigate fraud for a living or you just walked past someone wearing sunglasses on the train this morning, this story touches you. Your face is already out there — in photos you posted, in backgrounds of someone else's selfie, in databases you never signed up for. Meta's smart glasses look identical to regular Ray-Bans. A person wearing them could scan thousands of faces in a single day, and there's no practical way for a bystander to consent or even know it's happening. So the question running through this entire episode is simple. Where does a legitimate tool for solving crimes end and ambient public surveillance begin?

That R.S.A.C. demonstration is the right place to start, because it exposed a gap Meta hasn't addressed. Right now, Meta's glasses can't identify someone unless that person has a Meta account. The company built a wall between live identification and the broader universe of faces online. But that wall crumbled in a live demo. A researcher connected the glasses to an outside facial recognition service and suddenly had real-time stranger identification — names, profiles, all of it — running through hardware you can buy at a mall. The glasses didn't need Meta's permission. They just needed a camera and an internet connection.

That matters because it collapses a distinction most people have never thought about. Facial recognition and facial comparison sound like the same thing, but they aren't. Facial comparison is what a trained investigator does — taking an evidence image from a crime scene and carefully matching it against a known photo, with a human reviewing every result before anything moves forward. That human-in-the-loop step is what separates case work from mass scanning. Facial recognition in the way Meta's glasses enable it is the opposite. It's ambient. It's continuous. It runs on everyone in range whether they're a subject of an investigation or just grabbing coffee.

For anyone who works cases involving digital

For anyone who works cases involving digital evidence, that distinction is more than academic — it's their legal and ethical foundation. For everyone else, it means the person sitting across from you on the bus could be running your face through a search engine for humans without you ever knowing.

Meta pitched its planned feature — called Name Tag — as an accessibility tool for blind and visually impaired users. And that's a real use case with real value. But a leaked internal memo told a different story. According to that memo, Meta planned to launch the feature during what it called a "dynamic political environment" — a time when civil society groups would have their resources focused on other concerns. In plain language, the company expected backlash and tried to time its release for when watchdogs would be looking the other way. That's not a thoughtful rollout. That's a calculated bet that nobody would be paying attention.

Three U.S. senators — Markey, Wyden, and Merkley — have already demanded transparency from Meta about its facial recognition plans for these glasses. The coalition letter goes further, arguing that no amount of tweaking can fix a product that turns every pair of sunglasses into a surveillance device. Meta's response has been to note that competitors already offer similar products and that it hasn't actually shipped facial recognition on Ray-Bans yet. Both things are true. Neither one addresses the core problem — that the hardware is already in people's hands, and third-party software can unlock the exact capability Meta says it hasn't released.