Google pushes privacy angle as Gemini-powered smartglasses debut starting at $395
First reported by Ars Technica ·
Your photos and videos taken with these smartglasses won't train AI models or be used for ads, but your voice commands might be.
Google has launched its latest smartglasses, a collaboration with Samsung, Gentle Monster, and Warby Parker, priced starting at $395. These Gemini-powered glasses aim to address privacy concerns that plagued previous iterations like Google Glass. The devices feature a 12MP camera with a prominent recording light, a physical off switch, and software restrictions, including no live streaming to social media and no real-time facial recognition. Google states that captured images and videos will not be saved to Gemini history or used for ad targeting, though voice interactions are not guaranteed the same level of privacy. The glasses will ship on December 10 and are compatible with specific iPhone and Android models, excluding most Samsung and other OEM devices at launch. Samsung provides the hardware, while Google handles the AI software, integrating Gemini for personalized assistance and accessibility features like Gemini Guided Vision.
The integration of Gemini AI into smartglasses signifies a significant step towards more seamless and context-aware personal computing, potentially blurring the lines between digital assistance and everyday perception. By leveraging Gemini's capabilities, these glasses can offer real-time information, guidance, and interaction based on the user's visual and auditory input. However, the success of this venture hinges on Gemini's current fallibility; AI errors in object recognition or task execution, amplified by the lack of a visual confirmation screen, could undermine user trust and the overall utility of the device.
Google's emphasis on privacy features, while a necessary response to past criticisms, highlights the ongoing tension between data collection for AI improvement and user data protection. The implemented safeguards, such as the recording indicator and remote disabling, aim to build confidence, but the distinction in data handling between visual capture and voice interactions raises questions about the scope of privacy guarantees. The limited device compatibility at launch also suggests a cautious rollout strategy, likely targeting early adopters and specific ecosystems before broader market penetration.
AI-written summary. May contain errors.