Meta’s smartglasses look like ordinary Ray-Bans. That’s exactly the problem, and in the age of AI, the stakes for children have never been higher.
A man sits next to you on a crowded train. He looks like every other bored commuter. But he’s not scrolling his own photos. He’s zooming in on pictures of the toddler across the aisle, a little girl who belongs to the family sitting two feet away. Nobody notices, because nobody thinks to look. That’s the quiet horror writer Gaby Hinsliff described after watching it happen in person, and she says it’s exactly the risk Meta’s new smartglasses have built into everyday life.
Meta’s VP of wearables recently told the BBC how much he loved using his smartglasses to film his daughter’s talent show, calling it a way to stay “head up and hands free” instead of watching her through a phone screen. It’s a sweet story. But the same feature that lets him capture that memory lets a stranger do the exact same thing to somebody else’s daughter, without her parents ever knowing. The glasses have a tiny light that blinks when they’re recording. Most people have no idea to look for it, and hackers have already figured out how to disable it.
A Warning Light Nobody Notices
Here’s the uncomfortable truth: smartglasses don’t look like cameras. They look like glasses. Reviewers who’ve tested Meta’s Ray-Bans say friends sitting across the table never noticed they were being filmed. Compare that to someone holding up a phone. You see it coming. With smart glasses, you don’t.
That difference matters more than it might seem. A phone camera is obvious. Smartglasses are designed to disappear. The public has already started calling them “pervert glasses,” a nickname that stuck after reports of men secretly filming women in public and posting the footage online, some of it tied to extortion attempts. Nurseries and schools in the UK have started banning the glasses from their property altogether because they can’t tell when a parent picking up their own child is quietly filming someone else’s.
And it isn’t just strangers on the street doing the filming. A class-action lawsuit filed in March 2026 accuses Meta and its partner Luxottica of routing footage, including deeply personal moments, to contractors overseas for review without telling users clearly what was happening. Reporters found that workers reviewing that footage saw everything from people getting dressed to people using the bathroom. If a company can’t be trusted to handle footage responsibly on the back end, that’s one more reason to worry about what happens to images of children caught on the front end.
Meta has since added a feature that shuts off the camera if someone tampers with the recording light, and says privacy is “built in from the ground up”. But at the very same time, reports surfaced that the company is testing a prototype called “super-sensing” glasses, which would take photos every few seconds and record audio continuously, all without necessarily lighting up that warning indicator at all. One security lawyer summed up the contradiction bluntly, saying it looks a lot like a company saying “don’t look at the fire, look over here”. It’s hard to take a safety promise seriously when the next product in the pipeline may erase the one safeguard that exists today.
From Snapshot to Nightmare: The AI Problem
This is where the danger stops being about embarrassment and starts being about something much darker. A photo of a child, taken in public without anyone’s knowledge, can now be fed into an AI “nudify” tool and turned into fake but realistic child sexual abuse material in minutes. Some of these AI tools need only about 20 real photos of a child’s face to generate convincing fake abuse images. Right now, AI software can take any facial picture and turn it into anything you want, including videos that are pornographic in nature or highly disturbing. There are millions of people out there looking for this type of material, and it is a lucrative business.
The numbers back up how fast this problem is growing. The Internet Watch Foundation, a UK-based group that tracks child abuse content online, found that AI-generated videos showing child sexual abuse jumped by more than 26,000 percent in a single year, and nearly two-thirds of that material was classified in the most extreme category the organization tracks. In the United States, the National Center for Missing and Exploited Children received more than a million reports connected to AI-generated abuse images in just nine months, and federal investigators reported a 600 percent jump in AI-linked exploitation cases over that same stretch.
None of that requires a criminal mastermind. It just requires a photo, and smartglasses make it far easier to take one without anyone noticing. That’s why more than 70 civil liberties, domestic violence, and child advocacy groups sent a joint letter warning that Meta’s planned facial recognition feature for its glasses, which would let a wearer identify a stranger by name in real time, poses a “grave risk to privacy, safety, and civil liberties”. A UK charity focused on domestic abuse went further, warning the feature “threatens the safety of all women and girls in public”.
What Parents, and Lawmakers, Can Actually Do
Meta’s official position is that it’s up to individual users not to “actively exploit” the technology. That’s a little like handing out matches and asking people not to start fires. Digital rights groups aren’t buying it either. The Electronic Frontier Foundation has published a plain warning to think twice before buying or using these glasses at all, and recommends that current owners turn off cloud media sharing in the settings and stay alert to how footage gets stored and reviewed.
Some advocacy groups are pushing businesses, schools, churches, and hospitals to post signs banning the glasses outright, the same way many now ban weapons or outside food. Privacy experts have floated tougher fixes too: laws that force manufacturers to make recording obvious, maybe with an audible beep instead of just a blinking light; holding manufacturers rather than just wearers legally responsible when the glasses are used to cause harm; or requiring smartglasses to pass a safety review before hitting the market, the same way new drugs do.
Child safety expert Derek Baird has proposed one fix worth taking seriously. When Apple’s AirTags started turning up in stalking cases, Apple and Google built a shared system so any phone, iPhone or Android, warns you when an unfamiliar tracker is moving along with you. Baird argues smartglasses need the same kind of built-in alert: a signal that tells you when a recording device is nearby, restoring the awareness that a hidden lens takes away. It is not a perfect fix, since it warns bystanders without stopping the photo from being taken, but paired with a real effort to blur children’s faces automatically, it is a lot better than a blinking light nobody notices.
None of these fixes are perfect, and all of them would take real political will to put into place. In the meantime, the plain advice for parents is simpler: assume that any stranger wearing sleek sunglasses in a public place, a store, a playground, a school pickup line, might be recording, and that the recording could end up somewhere you never intended. Ask questions. Ask businesses your family visits to post a policy. And don’t wait for Meta to decide your child’s safety is worth more than a “never miss a moment” slogan.
The technology industry has a long history of building first and worrying about the damage later. This time, the damage lands on children who never got a say in the matter.