So Meta’s smart glasses have been recording people without anyone asking them to. No wake word, no permission, just… listening. And yeah, the company admits it happened, says it’s fixed now, but won’t actually explain what went wrong. Cool, cool. Nothing concerning about that at all.

When Your Glasses Become Informants

Here’s what went down: A user on Reddit noticed their Ray-Ban Meta glasses were picking up regular conversations – not the ones where they said “Hey Meta,” just normal talking-to-another-human conversations. And these weren’t getting lost in the ether. They were showing up in the Meta AI app as full audio recordings and text transcripts. The recordings just sitting there, archived, ready to be reviewed.

Now look, I get it. Tech has bugs. Updates break things. But this particular bug is the kind that makes your skin crawl a little bit, you know? Because the person who discovered this wasn’t just dealing with one faulty pair. They tested three different pairs – including both 2nd and 3rd generation models – and they all did the same thing. All of them started recording after a firmware update rolled out.

The white notification light would sometimes flick on when they were just talking normally. Sometimes it wouldn’t. Which is almost worse, honestly – at least with the light you’d know you were being recorded. Without it? You’re just walking around having private conversations that are being captured and transcribed without your knowledge.

The “Respond Without Hey Meta” Red Herring

And before anyone jumps in with “well did you check your settings?” – yeah, they did. The “Respond without Hey Meta” feature was turned off. Meaning the glasses should have been completely dormant unless explicitly activated. They weren’t supposed to be listening at all, let alone recording and saving everything.

What’s Meta Saying About All This?

Meta’s response has been… let’s call it minimal. They’ve confirmed the issue existed, said it’s been fixed, and basically left it at that. No deep dive into what caused it. No technical explanation. No transparency about how many users were affected or how long this was happening before someone noticed.

Which, I mean, come on. This isn’t a bug that made your battery drain faster or caused the app to crash occasionally. This is a bug that turned a piece of hardware people wear on their faces into an always-on surveillance device. That warrants more than “oops, our bad, it’s fixed now.”

The thing is, Meta’s glasses already had a trust problem. People were (and still are) worried about being recorded in public without their knowledge. The company tried to address this with the white LED indicator light – when it’s on, you know the glasses are recording. Except in this case, the light was inconsistent at best. So the one safeguard meant to signal “hey, I’m recording right now” just… didn’t always work.

This Keeps Happening

I’ve covered tech long enough to see this pattern over and over. Company releases wearable device with cameras and microphones. Company promises robust privacy protections. Device has “bug” that undermines those protections. Company issues vague statement. Rinse, repeat.

But here’s what really gets me: firmware updates are supposed to make things better, more secure, more reliable. This update apparently made the glasses start recording when they absolutely shouldn’t have been. How does that even happen? What was in that update that broke such a fundamental privacy control? Microsoft's AI Doom Loop: How It Could Kill the Internet explores what happens when tech companies prioritize features over the downstream consequences of their updates.

And we don’t know. Because Meta hasn’t told us.

“The conversations appeared in the Meta AI app as saved audio recordings and text transcripts.”

Let that sink in for a second. Not just recorded – transcribed. Processed. Stored. Someone’s private conversations turned into searchable text and sitting on Meta’s servers (or in their app, at minimum). Without consent. Without the wake word. Without any indication half the time that it was even happening. The AI Ad That Changed a Man's Race Without Asking shows what happens when tech companies process people's data in ways they never agreed to.

What This Actually Means

Look, I’m not going to sit here and say Meta did this on purpose. Bugs happen, even bad ones. But the lack of transparency around what caused this is a problem. Users deserve to know if this was a one-time glitch, a systemic flaw in how the wake word detection works, or something else entirely.

And honestly? This is exactly why people are skeptical of putting cameras and microphones on their bodies. Because when things go wrong – and they do go wrong – you’re not just dealing with a frozen screen or a lost file. You’re dealing with your privacy being compromised in real time, possibly without you even knowing it.

Meta says it’s fixed. Great. But until they actually explain what happened, how it happened, and what they’re doing to make sure it doesn’t happen again, those glasses are going to have a trust problem. And maybe they should.