// 9TO5GOOGLE — MOBILE & WEB
Here’s how Google’s Android XR glasses are handling the camera & privacy
Google’s Intelligent Eyewear push with the first Android XR glasses from Warby Parker and Gentle Monster are forced into an unfortunate uphill battle. The “perv glasses” stigma is very real because, frankly, early movers in this category didn’t respect privacy. Here’s how Android XR and Gemini are handling the camera and privacy on “Intelligent Eyewear,” and why I think Google has struck a near-perfect balance.
To quickly summarize the problem at hand (and name names), Meta’s smart glasses developed a reputation of being abused by bad actors who would disable the visible privacy LED (which Meta only recently fixed) or use these glasses in tasteless ways. Combine that with improper handling of recordings on Meta’s part and it’s a recipe for distrust and an earned social stigma. It’s not like Google is a simple bystander here, either. The original Google Glass project was met with the “Glasshole” stigma, and it’s something that Google has clearly learned a lot of lessons from.
It’s a shame, because a camera in smart glasses serves such a functional role, both for the user in just capturing pictures and for the AI assistant built into the experience. I recently talked about how the camera was my favorite part of smart glasses in an issue of our 9to5Google Weekender newsletter, but how it just feels a bit wrong now.
Anyway, with that context in mind, let’s look at how Google is handling this touchy subject.
It starts with the outward-facing light on every part of Android XR glasses, which works similarly to the privacy LED on Meta’s glasses. The key distinction here is this is a “recording light” on Android XR glasses. It turns on whenever the glasses are capturing content that can be viewed later in some form or fashion by the user. So if you’re taking a picture or recording a video, it lights up. If you’re just using Gemini for context, the light doesn’t activate. That’s because any camera captures being used for Gemini are deleted immediately after use. They won’t show up in your Gemini chat logs and that’s not a choice for the user – these images are purged immediately after they’ve been used so that you as the user nor Google can use them. That includes Google not using visuals from Gemini captures for AI training or advertising. Users also have control over what apps on Android XR glasses can use the camera.
Importantly, the light on each pair of Android XR glasses is also designed to avoid any tampering. Android XR glasses automatically perform a system check to ensure the light is working on a regular basis – you might catch the light blinking in white and blue while charging which is a part of the system check ensuring the hardware is fully functional. If the light is found not to be working, the camera is disabled.
One of the pieces of this package that really caught my attention is how Google is handling the privacy of those around you. When your Intelligent Eyewear glasses send a photo to Gemini, your phone automatically redacts faces from the image before it’s sent to Gemini. This happens locally on the phone, essentially as a part of the glasses app. Similarly, voice data sent to Gemini filters out other voices besides your own. The voice filtering bit is just one way in which Google is doing some impressive things with the microphones here, with another example being that the live translation feature works in a 60-degree “cone” in front of you, filtering out other voices in the space.
Continuing on with how the camera interacts with Gemini, Google explained in a conversation with 9to5Google that Gemini is tuned not to use the camera in cases where it’s not obvious that the user wants the camera to be active. An interaction such as “what painting is this” would very clearly imply that the user wants the glasses to view the scene in front of them and identify the art piece, where “how much money is on the table,” Google explained, would ask for a follow-up before using the camera seeing as it’s not obvious that