Google's Intelligent Eyewear separates Gemini visuals from saved photos in its privacy rules

Google says camera visuals used to answer questions through its new Intelligent Eyewear are not saved in Gemini History or used for model training or ad targeting. Photos and videos deliberately captured for a gallery follow a different path, so the promise should not be read as “the glasses never save images.”

Google detailed these protections on November 12, 2026, the same day as the originating WIRED report and the opening of preorders. Its collection announcement sets December 10 as the retail availability date for the eyewear developed with Samsung, Warby Parker and Gentle Monster. These are launch commitments, rather than results from TechKili testing.

A gallery photo and a Gemini question have different destinations

The Google Safety Center says a photo or video taken for the gallery goes to the paired phone. Users can choose to back it up elsewhere, such as a photo-storage app. That copy is separate from Google's statement about camera visuals used by Gemini.

For gallery capture, Google describes an outward indicator light and sensors designed to prevent capture, or stop a recording already under way, when the light is blocked. It also describes defenses against physical tampering. These are the manufacturer's design claims; we have not independently verified how reliably they work in different conditions.

For questions about the scene, Google's privacy announcement says the glasses are designed to redact bystanders' faces before Gemini processes the visuals. Face redaction is a particular protection, not a claim that every identifying detail in a room, document or conversation disappears.

The published indicator description specifically concerns photos and videos for the gallery. It should not be treated as a blanket explanation of every camera-assisted AI interaction. Someone deciding whether the glasses are acceptable in a shared space needs to understand the activity, not just recognize the frame.

Visual-data promises do not replace history settings

Google separately says chats are saved in Gemini History, with controls for reviewing and deleting conversations and for whether chat text improves generative models. Keeping camera visuals out of that history therefore does not mean the whole conversation has no retained record.

The Safety Center describes app-by-app camera and microphone permissions and a physical power switch that shuts both sensors down. These are useful controls to inspect during setup. They also give the wearer a concrete option when people nearby are uncomfortable: switch the electronics off and use the frames as ordinary glasses.

Our advice is to wait for independent testing if indicator reliability or redaction performance will determine whether you can use the device. Google's documented requirements already ask wearers not to record people without permission or record in sensitive spaces. A protective feature supports that responsibility; it does not establish permission on its own.

Sources

• Google's November 12 privacy announcement

• Intelligent Eyewear Safety Center

• Collection announcement and retail timetable

• Original WIRED coverage, November 12