The key facts
- Meta is reportedly testing "super-sensing" AI glasses that continuously record audio and snap photos every few seconds.
- To combat "LED fatigue," executives plan to disable the recording light during always-on AI monitoring.
- This follows a recent Meta update meant to stop creeps from secretly disabling the exact same light.
Literally 24 hours ago, I wrote an article praising Meta for pushing an emergency update to its smart glasses that disables the camera if someone tampers with the recording LED. I thought Zuckerberg was finally taking the peeping tom problem seriously.
Turns out I was an idiot.
According to a new report from the Financial Times, Meta is currently testing “super sensing” AI glasses that will continuously record your life — taking photos every few seconds and constantly listening to your audio. And the grand solution for the recording LED that warns bystanders? They’re just going to turn it off.
And with this new bombshell, it seems like Meta’s lost the privacy plot. The company didn’t build a software lock to keep the creeps out, they built one so they could be the only ones in the house.
‘LED fatigue’
The FT report notes that execs plan to not activate the LED when these super-sensing features are in use. Why? Well, in a 2025 policy paper, Meta laid the groundwork for this: “If the LED blinked for extended periods at a time… people could stop noticing it.”
For all the logic I’ve heard in my many years in this job, this is simply phenomenal gymnastics. “Your honor, I realized the fire alarm was annoying everyone, so I simply turned it off while the building burned down. You’re welcome,” says the lawyer in my head
The idea is that continuous recording will allow AI to help you “query what you saw or heard, or recall your day.” But to do that, it has to wiretap every single person you interact with, which is super illegal in multiple U.S. states (and heavily regulated in the U.K.).
The software update threat
Prototype-aside, here’s the scariest line in the FT report: “The super-sensing features could also be activated on Meta’s existing glasses via a software update.”
Yesterday, I was relieved that Meta built a software pipeline to stop a cottage industry of hackers breeding a new wave of creeps. Today, I’m realizing that the same pipeline could turn my glasses on my desk into a 24/7 surveillance device. There’s a big difference between words and action, and for all their talk of privacy, Meta doesn’t seem to be fixing the loophole; they seem to be just monopolizing it.
The proposed system from Meta argues that the raw content wouldn’t be stored. Instead, the metadata from the audio and images would be extracted and uploaded to the company’s servers for the AI to query. Some argue that it has “fewer privacy implications” — I see tech-bro speak for: "we aren’t keeping your recordings, we’re just taking a highly detailed, AI-generated transcript of everything you said and did."
Something has to change
Several people told FT that “plans could still change” around whether the LED is activated or not. But the fact that this is a serious thought Meta is having after all the backlash is kind of crazy.
It also raises an interesting legal question. We already know that super-sensing would raise all kinds of legal red flags — specifically, recording audio of a third party without their consent. But in that situation, are you (the wearer) committing a crime by walking into a coffee shop with these on? Or is Meta?
Woodrow Hartzog, a professor of law at Boston University, is absolutely talking my language when he told the FT that lawmakers need to update regulations for the reality of “always-on, always-seeing devices.” He’s absolutely on the money here. “There’s no one law that addresses all the different dangerous ways these tools have been designed and built,” Hartzog commented.
Yesterday, I wrote that true privacy protection will only happen when society decides that secret recording is socially unacceptable, and that we need to treat the cottage industry of Etsy modders who disable these lights like wiretappers.
It turns out we shouldn’t just be worried about the modders. We need to be worried about the trillion-dollar company too, because exactly as I said, relying on them to police themselves “never quite works out the way you think.”
I just didn’t expect that to be proven right within the space of a single day.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
- I’ve traveled over 50,000 miles testing smart glasses in 2026 — here are my 3 favorites that have launched this year so far
- I already paid $300 for Meta’s smart glasses — now it's charging $20 a month for a key feature
- Snap Specs vs Xreal Aura: The $1,000+ smart glasses that are fighting for your face this fall
