Meta's 'Privacy' Light on Its AI Glasses Doesn't Work

Going Viral Without a Say

Toluwa Omitowoju found out she'd been secretly filmed the way nobody wants to find out anything: friends, acquaintances, and people she hadn't spoken to in years started sending her a video of herself posted to social media. She had no idea she was being recorded. She wasn't asked. She had no say in what happened next.

The video racked up over 200,000 views across platforms before it was taken down, according to CBS News. Omitowoju was one victim in what CBS News describes as a growing trend of content creators using wearable AI devices to secretly record women for so-called prank pickup line videos. She says other women have reached out to her with worse experiences, including cases where the victims couldn't even identify who filmed them because the account posting the content was anonymous.

"It was a lot worse," she told CBS News. "It was to the point where they don't even know who filmed them and yet it's being posted online under a burner account."

This is the world Meta's Ray-Ban smart glasses are being deployed in. Not by tech journalists reviewing the specs. By guys making viral content out of women who didn't consent to be in it.

The Safety Feature That Isn't

Meta has a ready answer for privacy concerns about its glasses. The devices have a built-in LED light that activates whenever recording is in progress, warning anyone nearby that a camera is running. If someone covers that light, the glasses are supposed to detect the obstruction and shut off recording automatically. A neat little solution to a serious problem.

Except CBS News reporter Anna Schecter actually tested it, and the disabling mechanism did not appear to work when she covered the light after recording had already started. That is a significant caveat, and Meta's response to it tells you everything. A company spokesperson told CBS News the tamper detection only kicks in if the light is covered before you press record. Cover it after you've already started filming and the safeguard does nothing.

So the safety feature works in the one scenario where someone is planning, in advance, to be covert about it. If someone impulsively covers the light mid-conversation, or covers it after already capturing the footage they wanted, the glasses keep rolling. Meta told CBS News it "continues to work to improve this technology." Which is a very polished way of saying the thing they shipped does not do what people think it does.

Texas Is Asking Questions. Meta Has Answers Ready.

This isn't the first time these glasses have raised alarms. Back in May, Texas Attorney General Ken Paxton launched a formal investigation into Ray-Ban Meta glasses over allegations that the devices could be used to covertly film people, including in private settings. CBS News reports the investigation covered scenarios as extreme as bathroom visits.

Meta's response to Paxton's probe was a statement so smooth it practically has its own PR team. "Privacy and data protection are core to every product we build at Meta, including Ray-Ban Meta glasses," the company said. They said the same category of thing to CBS News when the LED light failure came up. Meta is very good at issuing confident-sounding statements about privacy. Building products that actually protect it appears to be the harder part.

The Law Is Not Going to Save You

Here is the part that makes everything worse. Even setting aside whether Meta's glasses work as advertised, the legal framework surrounding covert recording in public is almost entirely useless to the people being filmed. CBS News points out that most U.S. states operate under one-party consent laws, meaning only the person doing the recording has to know it's happening. The woman being filmed at a coffee shop doesn't count as a party whose consent matters.

Instagram head Adam Mosseri told CBS News in July that the platform is cracking down on this content, specifically calling out the pickup line video genre. "If you're posting content that is taking advantage of people and harassing them," he said, "then we're going to take the content down. We don't want people to be surreptitiously taking videos of other people and harassing them and then posting them on our platform." Fine. Good. But the content already exists. The views already happened. Omitowoju's face already traveled to 200,000 strangers.

Take-downs after the fact are not a privacy policy. They're a cleanup crew. And Omitowoju put it plainly to CBS News: with a phone, you can see it coming and object. With smart glasses, that option disappears entirely. "You're taking away the right of refusal," she said. There's no reclaiming that once the moment has passed.