Posts / privacy

The Little Light on the Glasses


The Dutch data protection authority has told people not to buy smart glasses. I read the headline, nodded, and then spent an hour reading the comments, which is where the interesting part was.

The advice is simple enough. These things film and listen, they send data somewhere, and the people around the wearer never agreed to any of it. The regulator also flagged some uglier details, and a commenter summed them up well. On some models you can put a bit of tape over the recording LED and the camera keeps working. That light is the only signal a stranger gets that they’re being filmed. If tape defeats it, the consent model is a sticker.

The other detail was security. On some pairs, a person nearby could apparently connect without permission, browse recordings, or run the camera remotely. So the risk isn’t only to the people being looked at. It’s also to the person wearing them.

I work in IT, so my first reaction was a professional one. A device with a camera, a microphone, a radio and an AI model attached will have bugs. A device sold as a fashion accessory will have more. I’ve seen enough poorly secured IoT kit to know that “the light can’t be disabled” usually means “we haven’t tested enough people with tape yet.”

What stuck with me, though, was a smaller thing. Someone described standing in a shop queue, chatting to the people around them. They turned, saw the bloke behind them was wearing smart glasses, and went quiet. They avoided looking at him for the rest of the wait. Another commenter replied, fairly, that if you’ve been inside a shop you’re already in a database. That’s true. Shops have cameras. Phones track location. Nobody reading this is a ghost.

But I think the commenter who went quiet was reacting to something real. A shop camera is bolted to the ceiling, has a sign, and belongs to an organisation you can complain to. A pair of glasses is a person. It’s a social space, and the camera has been moved into it. You can’t read the mood of a conversation the same way when you suspect one party is filing it somewhere.

Someone else said they’ve trained themselves to look for a camera on anyone wearing glasses, and keep the interaction short if there is one. That’s a bit sad. It’s also what happened to Google Glass a decade ago. People didn’t need a regulation. They just made it socially expensive to wear one. I’m not sure that works the second time, because these ones look like ordinary Ray-Bans. Glass was ridiculous on purpose, in effect. This generation is designed to disappear.

There’s a tension here I don’t want to wriggle out of. One commenter said they’d wanted good AR lenses for years to help with sensory processing, and now the whole category is tainted. Another pointed out that blind and low-vision users get real value from these devices, which is exactly why it matters when the AI gets medications or emergencies wrong. I believe both things. Assistive technology is among the best uses of this stuff, and it’s being carried in on the back of a consumer product that normalises covert recording. I don’t know how you separate the two. Maybe you don’t, and maybe you need a regulated assistive line with proper rules, which is dull and sensible and therefore unlikely.

I’ll also admit I’m not immune. I’m an iPhone person, I like gadgets, and when I first saw the demo videos I thought it looked clever. Then I pictured my teenage daughter at a bus stop in Cranbourne, and the clever feeling went away. She doesn’t get a vote on whether a stranger’s glasses are running face recognition, and neither do her friends.

Australia is behind Europe on this. Our privacy law is overdue for reform, and we’re mostly waiting for the next breach to prompt a response. A regulator saying plainly “don’t buy these” is more than I can imagine happening here without three inquiries first. I’d like to be wrong.

For now I’m not buying a pair. If someone I know turns up wearing them, I’ll ask politely about it. It might make for an awkward barbecue. That seems an acceptable price.