Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Waymo appears to be testing the addition of Google’s Gemini AI chatbot to its robotaxis with the aim of integrating an AI assistant that would accompany riders and respond to their queries, according to findings from researcher Jane Manchun Wong.
“Digging into Waymo’s mobile app code, I discovered the full system prompt for its never-before-seen Gemini integration,” Wong wrote in a blog post. “The document, internally titled ‘Waymo Ride Assistant Meta-Prompt,’ is a 1,200+ line specification that defines exactly how the AI assistant is supposed to behave inside a Waymo vehicle.”
The feature hasn’t shipped in public releases, but Wong says the system prompt makes it clear that it’s “more than just a chatbot.” The assistant would have the ability to answer questions, manage certain cabin functions such as air conditioning and, if necessary, reassure passengers.
“While we don’t have any details to share today, our team is still tinkering with features to make driving with Waymo enjoyable, seamless and useful,” Waymo spokesperson Julia Ilina told TechCrunch. “Some of these may or may not be part of our pilot experience.”
This wouldn’t be the first time Gemini has been integrated into the Alphabet-owned autonomous driving company’s stack. Waymo says it used Gemini’s “knowledge of the world” to train its autonomous vehicles to navigate complex, rare and high-stakes scenarios.
Wong writes that the assistant should have a clear identity and purpose: “a friendly and helpful AI companion integrated into a Waymo autonomous vehicle” whose primary goal is “to enhance the driver experience by providing useful information and assistance in a safe, reassuring and unobtrusive manner.” The robot should use clear, simple language and avoid technical jargon, and should keep its responses succinct to one to three sentences.
Depending on system prompts, when a pilot activates the assistant via the onboard display, Gemini can choose from a set of pre-approved greetings personalized with the pilot’s first name. The system can also access contextual data about the passenger, such as the number of Waymo trips they have taken.
Techcrunch event
San Francisco
|
October 13-15, 2026
Prompts currently allow Gemini to access and control onboard features, such as temperature, lighting, and music. Notably absent from the list of functions are volume control, route changes, seat adjustment and window control, Wong pointed out. If a pilot requests a feature that Gemini can’t control, the robot must respond with “ambitious phrases,” such as “That’s not something I can do at the moment.”
Interestingly, the assistant is responsible for maintaining a clear distinction between its identity as Gemini, the AI robot, and the autonomous driving technology (the Waymo Driver). So when you answer a question like “How do you see the road?” » Gemini should not say “I use a combination of sensors” and should instead answer: “The Waymo driver uses a combination of sensors…”
The system’s prompts include a range of compelling information, such as how the robot is supposed to handle questions asked about competitors like Tesla or the now-defunct Cruise, or trigger keywords that will stop it from speaking.
The assistant must also avoid speculating, explaining, confirming, denying or commenting on real-time driving actions or specific driving events. So if a passenger asks about a video they saw of a Waymo hitting something, the robot is asked not to answer directly and to deflect.
“Your role is not to be a spokesperson for the performance of the driving system, and you should not adopt a defensive or apologetic tone,” the message said.
The in-car assistant is allowed to answer general knowledge questions such as the weather, the height of the Eiffel Tower, the closing time of the local Trader Joe’s and who won the last World Series. Taking real-world actions like ordering food, making reservations, or handling emergencies is not allowed.
Waymo isn’t the only company integrating AI assistants into driverless vehicles. Tesla does something similar with xAI’s Grok. The two different car assistants, however, perform different functions. Gemini seems to be programmed to be more pragmatic and driving-oriented, while Grok is presented more as a traveling companion who can handle long conversations and remember the context of previous questions.
