Google details privacy protections in its new intelligent glasses

Google has set out the privacy protections built into its new intelligent eyewear, in a company blog post. Google says its research found wearers worried about using such glasses in public and bystanders worried about being filmed without knowing it.

When a wearer asks Gemini about something they see, such as translating a menu, the camera images are not saved in Gemini History and are not used to train AI models or target ads. Data on the glasses is encrypted, as is data sent to the phone or cloud for processing, and Google says it will soon bring its Private AI Compute technology to the glasses. Apps must ask for permission before using the camera for the first time, and access can be revoked per app.

Wearers can review and delete their Gemini chats and choose whether chat text is used to improve AI models. Every pair also has a physical power switch to turn the glasses fully off.