Gemini Live Guided Vision turns Android cameras into spoken visual assistants

Google is expanding Gemini Live with Guided Vision, a feature that uses the phone's camera to provide spoken information about objects and surroundings. A user can point the camera toward a label, sign or physical environment and receive guidance, while Gemini can request that the phone be moved when it needs a better view. The feature has obvious accessibility value for blind and low-vision users but can also help anyone interpret unfamiliar objects or hard-to-read information.

Google says support extends much further back through the Android installed base than many new AI features, reaching eligible devices from Android 9 onward in supported markets. That breadth matters because accessibility technology is most valuable when it is not confined to the newest flagship phones. Multimodal AI is often promoted through flashy demonstrations, but Guided Vision is a more meaningful test of reliability.

When software is being used to understand a user's physical surroundings, confident but incorrect answers can become genuinely problematic. Accuracy and clear uncertainty signals are therefore essential.