The camera sits at the corner of the frame on Ray-Ban Meta smart glasses, still visible today but harder to spot as styles diversify.
The camera sits at the corner of the frame on Ray-Ban Meta smart glasses, still visible today but harder to spot as styles diversify.
Meta AI glasses are the runaway success story of wearable technology, and that success has created a problem the hardware was never designed to solve. As millions of people buy a pair each year, the odds rise that the stranger beside you in a queue, a changing room or a doctor’s waiting room is wearing a camera you have not noticed and have not agreed to.
The only thing standing between a bystander and a silent recording is a small LED on the frame. Privacy researchers say it is not nearly enough, and a growing community of hobbyist programmers has stopped waiting for Meta to improve it.
The Warning Light Nobody Can Rely On
Meta AI glasses ship with a recording indicator built into the frame: a light that switches on when the camera is active. In practice, that safeguard is trivially defeated. A piece of tape or a sticker over the LED removes it entirely, and nothing about the glasses stops a wearer from doing exactly that.
Cooper Quintin, a staff technologist at the Electronic Frontier Foundation, described the feature to Ars Technica as nearly useless against anyone actually intent on misusing the device. The volume of unwanted recordings circulating online, he said, suggests the light is “better than nothing, but not by much.”
The documented misuse is not hypothetical. Reporting by 404 Media has covered the glasses being worn during an immigration raid and used to film and harass workers at their jobs. Elsewhere the pattern is grimmer and more domestic: partners and stalkers recording women in bathrooms and bedrooms. On social platforms, the same hardware turns up in everything from mild pranks to sustained bullying.
Where the Glasses Are Already Banned
Institutions have started drawing lines. Schools, courts, restaurants and entertainment venues have all introduced bans, each for slightly different reasons, and each running into the same enforcement problem.
The most striking prohibition came from the security community itself. DEF CON 2026, an event attended by people who build and break surveillance technology for a living, banned smart glasses outright with no exceptions, telling attendees who wear prescription pairs to “be sure to pack non-violating eyewear if you need them.” When the hacker conference decides a device cannot be trusted in the room, the signal is hard to miss.
A ban, though, is only as good as the person enforcing it. Spotting smart glasses takes a security guard who knows what to look for, and in most venues enforcement depends on wearers volunteering that they have a camera on their face, or on another guest noticing and speaking up.
What the Detector Apps Actually Detect
Since February, independent developers have been shipping phone apps that scan for nearby smart glasses, including Ray-Ban Meta, Oakley Meta and Snap Spectacles. The newest arrival, Zuckoff, is free. The most established, Nearby Glasses, costs $2.99 on iOS and Android and has passed 100,000 downloads on the Play Store alone.
What these tools actually do is narrower than most buyers assume. The apps listen for the Bluetooth signal that pairs a set of Meta AI glasses with a phone. That tells you a device is powered on and in range. It does not tell you whether the camera is rolling.
The developers are unusually candid about the limits. Yves Jeanrenaud, who wrote Nearby Glasses, warns on the project’s GitHub page that he is “no pro developer,” that Bluetooth LE detection “might sometimes just not work as expected,” and that “false positives are likely.” A wall or even a person standing in the way can block the signal. App makers advise against confronting anyone and disclaim liability for a misread.
None of that has dented demand. Users appear to want something simpler than certainty: a way to scan a room and know that a camera is present. Many send the developers feedback, and some describe an emotional investment in seeing the tools improve.
The Face Scanner That Reached 50 Million Phones
The reason privacy groups treat detection as urgent rather than paranoid is what sits one software update away. In June, Wired and the EFF revealed that an unreleased facial recognition system had been quietly installed inside Meta’s smart-glasses app, on an estimated 50 million phones. After the disclosure and the public reaction that followed, Meta deleted the code.
The EFF called the removal a victory, while noting it settles nothing permanently. There is no public commitment preventing the feature from returning, and the group argues that makers will keep reaching for facial recognition unless lawmakers give people a private right to sue over biometric privacy violations.
Quintin’s concern runs past the recording itself. “I think that there are a lot of other concerns there beyond just being recorded without your knowledge,” he said. Glasses that identify a face in real time, paired with generative tools, open the door to convincing deepfakes and intimate imagery of people who never knew a camera was pointed at them.
Super Sensing Would Turn the Light Off
There is also a live question about whether the indicator light survives at all. The Financial Times reported last month that Meta is testing a prototype of “super sensing” glasses that would use cameras and audio recording to capture a wearer’s every moment.
According to sources familiar with the company’s thinking, executives are planning not to activate the LED while super-sensing features are running, even though, as the report put it, “that would make it harder for bystanders to know when they were being recorded.” Those plans could shift if privacy concerns stay loud, the same sources suggested. Meta did not respond to a request for comment on the story.
That is the fault line. The one visible safeguard on the device is under review not because it works too poorly, but because it works at all.
What Bystanders Can Realistically Do
Quintin’s argument is that neither approach works alone. “You need to have the social construct around these being unwelcome, and you also need to be able to find when people are using these,” he said. Bans set the expectation; detection tests whether the expectation is being honoured.
The window for doing it by eye is closing. “Right now, they’re pretty easy to just detect with your eyes. They’re pretty obvious, but that might not always be the case,” Quintin said. Once the frames diversify into ordinary-looking styles, the visual tell disappears and software becomes the only check left.
For anyone weighing what to do now, the practical advice is unglamorous. Treat a detector app as a rough scan rather than proof. Never assume a person wearing glasses is recording you. Support venue bans where they exist, because enforcement depends on guests noticing. And if you wear a pair yourself, the EFF’s user guide asks for the obvious courtesies: sensible recording habits, an understanding of what continuous capture means, and blurring strangers’ faces before anything goes online.
The wider policy fight is the same one running through consent and identity rules across the internet, from biometric protections to the age-verification proposals now reshaping how people prove who they are online. Until comprehensive privacy law catches up with the hardware, the burden keeps landing on the person who did not buy the glasses.
This article was AI-assisted and edited for accuracy.
