LifePivo

Meta Hardened the "Pervert Glasses" Recording Light in July

A flat illustration in bright midday daylight: a large pair of chunky black-framed glasses fills the foreground on sunlit
AI-generated editorial illustration. Not a documentary image.

Explained by Maya Chen — LifePivo fictional editorial host

·

Sources last checked

the fix closed a hack, not the light no one notices

There is a small white light on the corner of Meta's smart glasses that is supposed to be the whole social contract. It blinks when the camera is running. That is the deal the glasses offer the strangers around you: you will know when you are on camera, because the light will tell you.

In July, Meta made that light much harder to defeat. A mandatory firmware update, dated 7 July on the company's own newsroom, now shuts the camera off entirely if the glasses detect that the light has been physically tampered with or destroyed: not merely covered. Covering it with tape had already killed the camera since the second-generation frames, because an ambient-light sensor sits behind the same window. The people who wanted a truly silent camera had moved on to drilling the LED out of the hardware, and a small cottage industry had grown up selling the service. Meta's update reaches back and disables those glasses too, locking the camera until the device can confirm the light works again, and the company says it is pulling the ads for LED-removal services, banning the accounts that sell them, and taking legal action against the sellers.

It is a real piece of engineering, and Meta is right that no ordinary camera polices itself this way. And it did almost nothing for the problem people are actually angry about.

The label stuck, and the fix did not shake it loose

A month after the update, the glasses were still being called "pervert glasses": in comment sections, in the press, and by the people who had agreed to promote them. The Guardian reported on 9 August that content creators who film with the glasses now draw a flood of hostile replies whatever they are filming. Rida Tharana, who has 1.6 million Instagram followers and had taken part in a Meta campaign, publicly "took accountability", turning on the men who use the glasses to "record and harass women". A parenting creator, Deena Lang, described watching back footage from a cooking class and being "horrified": not by anything she had done, but by the row of strangers in the frame giving her side-eye, none of whom had agreed to be there. Her glasses, she said, are now in a drawer.

None of that was a response to the drilled-out-LED exploit. These were ordinary glasses, worn as sold, light blinking as designed. The backlash is not about people who hacked the hardware. It is about people who did not.

That gap is the whole story, and it is worth being blunt about it: Meta answered the version of the problem it could engineer its way out of, and the version left standing is the one built into the product on purpose. A camera that looks exactly like a pair of glasses is a camera most people will never clock, and a blinking light the size of a grain of rice does not change that. Fortune, writing about the same update in July, made the point plainly: the light may not register with people who do not know what it means, and it is hard to see in daylight at all. The update did not touch either of those. It could not. The thing that makes the glasses uncomfortable is the thing that makes them sell.

And they sell. Meta and its eyewear partner EssilorLuxottica reported around seven million pairs of smart glasses in 2025, more than triple the year before, and this summer the line went openly mainstream: a June launch put Kylie Jenner's name on a slim oval frame, part of a range starting at $299, with the option to hear the assistant answer in her voice. Seven million pairs is seven million cameras that are indistinguishable, across a table, from a pair of reading glasses. The scale is the point, and it is also the reason a firmware patch aimed at a few thousand modified units misses the room.

What the light is actually up against

To see why the fix and the fear are aimed at different targets, it helps to separate the two ways these glasses can record someone who has not agreed to it.

The first is deliberate concealment — defeating the light so the camera runs invisibly. That is a hack, it is what July's update kills, and it was never how most footage got made, because it did not need to be. This is the same question our own close reading of whether you can tell the glasses are recording kept running into: the honest answer, even with the light working perfectly, is often no.

The second is the ordinary case, and it needs no modification at all. You wear the glasses in public, the light blinks as intended, and the people around you simply never notice — because it is small, because it is daytime, because they are not looking for it, because you look like someone wearing glasses. Everything the "pervert glasses" complaint is about lives in this second category, and July's update does not enter it. We went through exactly what the indicator does and does not cover in a separate piece on the recording light; the short version is that the light is a disclosure only to a person who already knows to read it.

Meta's own defence, given to the Guardian, leans hard on the first category and quietly steps around the second. Its glasses were designed "with privacy and safety features built in from the start", a spokesperson said, "informed by engagement with safety experts, including women's safety organisations", and the tamper detection means "no other kind of camera has done this". All true, and all about the version of the problem that involves someone deliberately breaking the light. Not one word of it answers the wearer whose light is working and whose subjects still had no idea.

The incidents that kept coming

The reason this is not an abstract complaint is that the concrete cases kept arriving, and they did not involve tampered hardware either.

The one that reached the most people was reported by the BBC: a woman in London said a stranger filmed her, posted it, and then told her that taking the video down was a "paid service" — recording turned into a small extortion racket. The footage reportedly drew tens of thousands of views; the man denied trying to extort her, and police said there was too little information to open a formal investigation. Then, in early August, Police Scotland confirmed it had received complaints and was reviewing footage taken by a TikToker who, over a road trip around Scotland, is alleged to have filmed members of the public, young women, shop staff, without their knowledge, and to have confronted a Sikh shopkeeper on Skye, a man who has lived on the island for twenty years, with a reference to a far-right conspiracy theory. The shopkeeper told reporters the encounter was "100% intimidating" and cried when he learned the video existed. Meta says it is investigating that user's account. No one in either case has been charged.

Hold those two next to the July update. In neither did the light need defeating. In neither would tamper detection have changed anything. The camera worked as designed; the disclosure the light is supposed to provide simply did not land, because a blinking dot on a stranger's face is not something a shopkeeper mid-conversation, or a woman walking through a shop, is watching for.

The honest read

So here is the sentence we are willing to be wrong about: the "pervert glasses" backlash will not fade because Meta hardened the recording light, because the light was never the part people distrust. What they distrust is a camera they cannot see being worn by someone they cannot read, and Meta has spent a year and seven million units making that camera more ordinary, not less. The update is a genuine improvement against a narrow abuse, and it is orthogonal to the thing driving people to put the glasses in a drawer.

There is a version of these glasses that could answer the fear: a larger, unmissable indicator, a louder signal, a design that reads as a camera rather than as spectacles. It would also sell worse, because looking like an ordinary pair of glasses is most of the pitch. That is the trade Meta has not made, and July's update, for all its cleverness, let the company look busy on privacy without going near it. The people calling them "pervert glasses" have, in their blunt way, understood the product better than the patch notes have.

Sources and verification

The product change is documented by Meta's own newsroom and corroborated across several independent tech outlets that quote it; the incidents are attributed to the outlets that reported them, and where a claim rests on a single report it is marked as such. Where a case involves a named individual, it is stated as an allegation under review, because no charge or conviction has been reported.

Wetherspoon and other UK venue restrictions are covered in our own report, are smart glasses banned in pubs and restaurants.

What we could not confirm: that the backlash is quantifiably larger than before (the dated incidents show it persisting after the fix, not a measured increase); that the July update was a response to the backlash rather than to misuse amid growing adoption, which is the reason Meta itself gives; and the precise unit total, which rests on the companies' own reporting.