Gemini Live Guided Vision expands camera-based assistance across Android

Google is expanding Gemini Live's Guided Vision capabilities across the Android ecosystem, allowing users to point a phone camera toward objects or surroundings and receive spoken assistance. The system can describe what it sees and instruct the user to reframe or move the camera when the subject is not positioned clearly enough. Accessibility is the most obvious application, particularly for blind and low-vision users who may need help reading labels, interpreting menus or locating household objects.

Google says Guided Vision will extend as far back as Android 9 on eligible devices, making it significantly more inclusive than flagship-only AI features. The technology also provides a useful test of multimodal AI reliability. A chatbot giving an imperfect answer to a trivia question is annoying; inaccurate guidance about the physical environment can be more consequential. Google's challenge is therefore not simply recognising more objects but communicating uncertainty clearly when the model is not confident.

If it works reliably, the camera becomes a conversational sensor rather than just a photography tool.