Google details privacy protections for Intelligent Eyewear

Google says its Intelligent Eyewear is being built with privacy controls aimed at both the person wearing the glasses and everyone around them. The company says Gemini camera visuals are not saved in Gemini History, used to train its AI models or used for ad targeting. Data stored on the glasses, or sent to a phone or cloud service for processing, is encrypted. ¹

That matters because the glasses are designed to look at the world around the wearer. Google says they can help with tasks such as translating signs, reading menus and navigating, while Guided Vision in Gemini Live can provide audio descriptions for people who are blind or have low vision. The more useful the camera becomes, the more important it is to explain what happens to the footage and who controls access to it.

What Google says the glasses protect

The company says each pair has an outward-facing indicator light that alerts nearby people when photos or video are captured. Google also says the glasses can detect attempts to tamper with that light and block capture. Faces of bystanders are automatically redacted before Gemini processes visual information, according to the company, while directional microphones are intended to focus on the wearer's voice.

Users can reportedly control which apps access the camera, revoke those permissions, manage Gemini History and delete individual chats or the full history. A physical power switch provides a more direct option: the glasses can be turned off rather than relying only on software settings.

The limits of the announcement

These are Google's stated protections, not independent proof that the system prevents every form of misuse. The practical effectiveness of face redaction, the reliability of tamper detection and the handling of unusual environments still need to be tested outside the company's own demonstrations. The distinction is especially important for wearable cameras, where bystanders may not know exactly when a device is processing what it sees.

Our opinion

Google is right to frame privacy as a two-sided problem rather than a setting that only protects the wearer. A visible capture signal, a physical power switch and user-controlled permissions are understandable safeguards. But understandable safeguards are not the same as a guarantee. The real test will be whether the indicator is consistently visible, whether redaction works quickly enough in messy real-world scenes and whether Google explains its retention policies with the same clarity it brings to the product pitch. For now, this is a useful account of the intended design — not a verdict on how well Intelligent Eyewear works in public.