Loading page…
Sales start December 10. Google promises face redaction and tamper-resistant recording lights, but privacy experts say safeguards do not settle consent.
Listen to this story
Google’s privacy design separates ordinary photo and video capture from Gemini’s visual assistance: the glasses use a recording light for capture and are designed to blur bystanders’ faces before AI processing. The controls do not settle whether people nearby have meaningfully consented, a concern raised by privacy advocates and researchers. Preorders are open, with sales due to begin December 10; Google has not given a release date for its planned Private AI Compute system.
Video capture is capped at 90 seconds, and Google says sensors can block recording if the indicator is covered or tampered with.
Camera visuals used for answers are excluded from Gemini History and, Google says, from AI training and ad targeting; chats are saved and have separate controls.
Directional microphones prioritize the wearer’s voice, while apps must request camera access and owners can revoke access individually.
Google’s first Android XR AI glasses are available for preorder, with sales starting December 10, according to WIRED. Alongside that release, Google detailed protections for wearers and people nearby on November 12. The safeguards address both recording and AI processing, but privacy experts argue that technical controls leave questions about bystander consent unresolved.
The Intelligent Eyewear glasses were developed with Samsung, Warby Parker and Gentle Monster. Google’s privacy announcement distinguishes gallery recording from letting Gemini examine what the wearer sees.
For gallery photos and videos, an outward-facing LED signals capture. Google says sensors are designed to prevent recording if someone covers or tampers with the indicator. Video recording is limited to 90 seconds for privacy purposes, WIRED reports.
Juston Payne, Google’s senior director of product management for XR, told WIRED that algorithms continually check LED functionality, acknowledging the system is not perfect.
Gemini’s visual assistance has a different protection: Google says the glasses are designed to redact bystanders’ faces before the AI processes images. Directional microphones focus on the wearer’s voice by default, with noise removal intended to keep nearby conversations from reaching Gemini.
Google says camera visuals used to answer questions are not saved in Gemini History or used for AI training or ad targeting. Chats, however, are saved for continuity and personalization. Users can delete them and control whether chat text helps improve generative AI models.
Data is encrypted on the glasses and in transit to a phone or cloud, Google says. Private AI Compute, intended to process data in a secure space, is planned but has no release date.
Apps must request camera access, which wearers can revoke individually. A physical power switch turns the device completely off for use as ordinary glasses.
Fight for the Future’s Reem Suleiman told WIRED that camera-equipped glasses should not become normal in public, citing surveillance risks and a lack of actual consent. Oxford researcher Janusz Swierczynski called the protections meaningful but said they largely address individual privacy rather than everyone affected. He warned that normalizing recording can have unintended consequences.
Loading discussion...
Join the conversation
Explain what would make you comfortable—or uncomfortable—around someone wearing them.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.