Gemini in the Car: Google Tests an Assistant That Can Look Ahead

Google has shown an experimental Gemini feature for cars: the assistant could temporarily access a vehicle’s external camera and answer questions about what is in front of the driver. The capability was demonstrated by Android president Sameer Samat during a conversation with Google AI Studio head Logan Kilpatrick.

The demo took place in a Volvo EX60, a model that has not yet gone on sale. The car runs on Google built-in with Gemini and, judging by the demo, currently appears to be the only vehicle where this scenario is available. Google has not said when the technology might reach other models, or whether it will at all.

The idea is similar to the camera mode in Gemini Live on Android phones, except the car camera becomes the assistant’s eyes. A user can ask Gemini about a building, object or sign ahead, and the system processes the image before replying by voice. Samat also stressed that Gemini does not analyze the road continuously: the camera connects only after an explicit assistant request and a specific question, then access is turned off again.

During the demonstration, Gemini recognized objects around Google’s Mountain View campus. The assistant identified local landmarks, including The Orb art installation, described details of the Gradient Canopy solar roof and explained the architectural history of Shoreline Amphitheatre. The system was not instant: there was a noticeable pause before the camera connected, and the car screen did not show a viewfinder like Gemini Live does on a phone.

Google still calls the feature experimental and is looking for the best real-world driving uses. One example is translating and explaining road signs in a foreign language while traveling. If the technology reaches a wider launch, Gemini in cars could answer not only using maps and navigation, but also with context from what the vehicle actually sees ahead. The original material did not mention pricing.