Waymo Testing Gemini AI Assistant to Transform Robotaxi Experience
Waymo is testing Google’s Gemini AI as an in-car assistant for its robotaxis. This new feature allows riders to control comfort settings like temperature and music via voice, enhancing the passenger experience while keeping driving controls separate for safety
Waymo has begun testing an in-car voice assistant powered by Google’s Gemini artificial intelligence in its fleet of autonomous vehicles. The integration, which was uncovered in the company's app code [December 24, 2025], aims to make ride-hailing more interactive by allowing passengers to control vehicle settings and ask general questions during their trips.
According to findings by app researcher Jane Manchun Wong, the new feature is designed to function as a "ride assistant" rather than a driver. The system prompts discovered in the application reveal that the AI can handle tasks such as adjusting the climate control, changing interior lighting, and managing music playback. However, it is explicitly restricted from accessing safety-critical functions; the assistant cannot alter the vehicle's route, open windows, or operate the doors.
The integration emphasizes a clear distinction between the conversational AI and the autonomous driving system, known as the "Waymo Driver." The leaked system guidelines instruct the Gemini assistant to avoid using first-person language when discussing driving maneuvers. For instance, if a passenger asks how the car sees the road, the assistant is programmed to reply that "the Waymo Driver uses a combination of sensors," rather than saying "I use sensors."
Waymo confirmed to TechCrunch that its teams are utilizing Gemini's capabilities to explore new features, though a full rollout has not been officially dated. "Our team is constantly working to improve the Waymo experience to make it comfortable, seamless, and convenient," a spokesperson stated.
The move comes as the autonomous vehicle sector shifts focus from purely technical capability to passenger experience. By utilizing Google's Gemini models, Waymo aims to provide concise, helpful responses, capped at three sentences, to reassure riders and reduce the silence often associated with driverless rides.
