Let’s start with what’s happening on the glasses. It’s basically table stakes at this point, but they all have an LED light on them to indicate when someone is taking a picture or recording a video. The light, which sits on the left side of the frame while the camera is next to the right eye, has sensors to detect if it’s covered or tampered with and will prevent the capture of pictures or video.
“Unfortunately, some of the glasses that have launched in the past launched without any kind of detection for those things, and that broke trust.” Payne said. “And that’s a problem,” he added. He also said Google’s methods are “pretty effective,” though that’s something that we’ll have to test for ourselves to verify.
And just as there are bad actors who will continue to try to break security protections for any safe in the world or online, there will be people attempting to get around Google’s anti-tampering measures here. Payne said “there will always be ways around what we build, and then we’ll try to fix those holes,” and that he expects there will be an ongoing back-and-forth. “But nonetheless, we’re very happy with where we’re starting on this one.”
One of the main functions of Google’s smart glasses is the ability for Gemini AI to use the camera to see what the wearer is looking at and provide more relevant information. During this process, the company says that the camera visuals won’t be stored in your Gemini history. When there are faces in view, Google says “your glasses are also designed to automatically redact bystanders’ faces before Gemini processes the visuals.”
The camera stream also won’t be used in AI training or ad targeting, and Google says any data that is sent off the glasses for processing, say in the cloud or on your phone, is encrypted for transit. It’s also promising to bring the Private AI Compute processing platform to glasses, enabling these new devices to use the “hardware-secured sealed cloud environment” for greater processing.
While the camera may be the most obvious recording sensor on these devices, they also have four microphones onboard to allow you to speak to Gemini or get translations for people around you. Google uses beamforming to help focus on the individuals you’re looking at when deciding whose speech to translate, and the company also said it “designed Gemini to focus on your voice by default, and not other voices around you, using directional microphones on the glasses.”
Payne said that the microphones will filter out “all the world audio, as we call it, the bystander audio,” It helps prevent accidental recording of people in your vicinity, but also improves Gemini’s effectiveness by making sure it doesn’t confuse someone else’s heated argument as your query.
Of course, to make the glasses more useful, you’ll have to grant it access to your apps and personal information. The intelligent eyewear can work with apps like Messages, Keep or YouTube Music, for instance, to send replies or create lists for you. Like on most phones, you’ll have to grant permission to specific apps before they can access the camera on your glasses, and even after you do, you can still add or revoke access for every app.
Google also said you can view your past interactions with Gemini, delete individual chats or all of them. You’ll also be able to decide if your conversations are used to train AI. And to make sure there’s no confusion over whether you left the glasses on or off, and if the microphones or camera might still be tuning in, all the new eyewear have a power switch with a clear red indicator to tell you when they’ve been shut down.
