Thursday, 12 November 2026

How to tell if Google glasses are recording, and when you can’t

A light on the left side of the frame marks every saved photo or clip, but Gemini can still look through the camera with that light off.

Smart glasses on a cafe table with a small indicator light glowing on the frame

The short version

  • Google's new Android XR glasses light up an outward-facing LED whenever they save a photo or video.
  • If the light is covered or broken, the glasses refuse to take photos or video, and video clips stop at 90 seconds.
  • When the wearer asks Gemini about the scene, the camera works with the light off, but Google says those images are deleted at once.
  • Faces are redacted before Gemini sees an image, and Google Play bans apps built for real-time face identification.
Quick read · 2 min

Google's first Android XR glasses, made by Samsung and sold through Warby Parker and Gentle Monster, light an LED on the left side of the frame whenever they save a photo or video. Cover or break that light and the camera won't capture anything. Videos are capped at 90 seconds, and a physical switch with a red indicator turns the whole thing off.

The catch: when the wearer asks Gemini about the scene, the camera can look without the light. Google says those images are deleted right away, never stored in Gemini history and never used for ads or AI training, and faces are redacted before Gemini sees them.

  • Google Play bans face-ID, live-streaming and light-changing apps.
  • Google Shopping and Ads block light-hiding stickers.
  • Wearers must agree not to film people without consent.

Preorders are open now from $395, with sales starting December 10.

A small light on the left side of the frame is the giveaway. On Google’s new Android XR glasses, that light switches on whenever the camera saves a photo or a video, and if someone tapes over it, the camera simply won’t take the shot. There is one exception, though, and it matters: when the wearer asks Gemini about what’s in front of them, the camera can look at the scene while the light stays dark.

That’s worth knowing now, because you’re about to see a lot more of these frames. The first Intelligent Eyewear models from Warby Parker and Gentle Monster, built by Samsung with Gemini inside, are on preorder at $395 and up, and sales start December 10. After a year in which camera glasses picked up nicknames like “creep glasses,” Google has laid out in unusual detail what its glasses record, what they throw away and what they refuse to do. Here’s how to read the signals, whether you’re wearing a pair or standing across the table from one.

01

How to tell if Google glasses are recording: watch the left side of the frame

The outward-facing light is your main clue to whether Google glasses are recording. Google calls it a recording light, and the distinction is deliberate: it comes on only when the glasses capture something the wearer can look at later, which means a photo or a video. The light sits on the left side of the frame, while the 12MP camera sits near the right eye, so look at the opposite side from the lens.

Three other details help you judge what’s going on:

  • Video is short by design. Each recording stops at 90 seconds. Google says most clips people shot already fit inside that limit. Meta’s glasses, by comparison, allow clips of up to 5 minutes.
  • There’s a real off switch. Every pair has a physical power switch that cuts everything, the microphones included, so the frames work as ordinary glasses. A red indicator shows when they’ve been shut down.
  • A blinking light on the charger is normal. If you catch the light flashing white and blue while the glasses charge, that’s a self-test checking the hardware, not a recording.
Eyeglasses held up in front of a blurred city street at dusk
02

What happens if someone covers the recording light

Covering the light stops the camera. Google says the glasses watch in real time for a finger, sticker or tape over the LED during a photo or video, and also check whether the light itself has been damaged. If either happens, capture is blocked. The glasses also run regular checks to confirm the light still works, and if it doesn’t, the camera is switched off.

Juston Payne, Google’s senior director of product management for XR, has been candid that this isn’t foolproof. He told reporters “it’s not perfect” and expects people to find workarounds that Google then has to close, an ongoing back-and-forth. That’s why the company added rules outside the hardware:

  • When you set up the glasses, you must agree not to hide the light, not to film people without their consent and not to record in sensitive places. Breaking those terms can get your glasses disabled.
  • Google Shopping and Google Ads will block sales and ads for stickers or covers made to hide recording lights.
  • Setting up a pair needs a Google account belonging to someone 18 or older.
03

When the camera looks without the light: Gemini requests

Yes, Google glasses can use the camera without the light, but only for Gemini. If the wearer asks Gemini to translate a menu or read a poster, the camera grabs what it needs and the light doesn’t come on. Google’s reasoning is that nothing from those requests is kept: the images are deleted right after use, don’t appear in the wearer’s Gemini history, and aren’t used for AI training or ad targeting. The wearer can’t change that.

Google adds several layers on top:

  • Faces are blurred out first. Before Gemini processes an image, the phone app redacts the faces of people in it. In one hands-on test, Gemini told a reviewer it could see a person but wouldn’t identify him, and it refused the same task with a printed photo. Payne says the step exists to defeat face recognition.
  • Gemini checks before looking. If a request doesn’t make it obvious that the wearer wants the camera, such as a question about money on a table, Gemini is tuned to ask a follow-up first. A clear request like “what painting is this” goes straight through.
  • Data is encrypted. Visual data is encrypted on the glasses and while it travels to the phone or the cloud. Google also plans to bring its Private AI Compute system, its sealed cloud setup, to the glasses, though that’s still listed as coming soon.

Audio gets similar treatment. The glasses carry four microphones that focus on the wearer’s voice by default, and other people’s conversations are filtered out before anything reaches Gemini. Live translation listens to roughly a 60-degree cone in front of the wearer.

Smart glasses and charging case beside a smartphone on a desk
04

What this means if you’re the one being filmed

For bystanders, the rule of thumb is simple. A lit LED on the left side means a photo or clip is being saved, and any clip will end within a minute and a half. No light means nothing is being saved, though Gemini may briefly be looking at the scene, with faces redacted, if the wearer has asked it something.

That still leaves gaps. Privacy experts point out that a small light on someone’s face is much easier to miss than a phone held up at arm’s length. Reem Suleiman of the digital rights group Fight for the Future argues that camera glasses shouldn’t become normal in public at all, calling them a step toward mass surveillance. Google’s promises about deleted images and blurred faces are policies you have to take on trust, and much of the protection only works because the light can’t be easily defeated.

If footage of you does end up online, YouTube’s rules against harassment and unwanted sexualization apply no matter what device shot it. Google says videos that break them are removed, and it plans to demonetize such content too.

05

The settings Google glasses wearers control

Wearers get phone-style controls over what the glasses can do:

  • Camera access per app. Third-party apps must ask before using the camera, and you can take that permission back at any time. Apps that want camera or microphone access face extra review.
  • Gemini history. You can look back at past Gemini chats, delete single ones or wipe them all, and choose whether your conversations help train AI.
  • The power switch. When you want the glasses to be just glasses, flip it off and the red indicator confirms it.

What wearers can’t do is just as telling. Google Play rules bar apps from live-streaming to social platforms, from identifying faces in real time and from changing how the recording light behaves.

06

How Google’s approach compares with Meta’s glasses

Meta’s glasses also have an LED and also stop capture when it’s blocked, but their track record is rougher. Bad actors found ways to disable Meta’s privacy light before the company fixed it, and this summer Meta quietly added face-recognition code to its glasses, then removed it after it was reported. Meta’s glasses also allow much longer videos.

Google’s bigger difference is reach. Because it runs an app store, an ad network, a shopping service and YouTube, it can go after light-hiding accessories, face-ID apps and harassing videos well beyond the glasses themselves. Whether that changes public opinion is the open question once the first pairs arrive on December 10.

07

Quick answers

Do Google smart glasses have a recording light?

Yes. Every Android XR pair from Warby Parker and Gentle Monster has an outward-facing LED on the left side of the frame that turns on for photos and videos.

Can you turn off the light on Google glasses?

No. Covering or damaging it blocks photos and video, Google Play bans apps that change it, and hiding it breaks Google’s terms.

Do Google glasses use facial recognition?

No. Faces are redacted before Gemini sees an image, and Google Play prohibits apps that identify people in real time.

You're reading the quick version.