Key Highlights:
- Waymo is testing Gemini integration for its robotaxis, according to Jane Manchun Wong, who is a tech blogger and app researcher.
- Gemini can do much more than just being a simple in-car assistant; possibilities include answering your queries during rides to managing certain in-cabin feature for making your ride seamless.
- It’s unclear when Waymo is planning to integrate Gemini into its robotaxis.
Google has been making massive strides to turn Gemini into a useful AI assistant that not only helps us with summarizing emails, generating images, and much more. Thanks to the growing list of features now associated with Gemini, Waymo appears to be exploring a new frontier for its robotaxis. Formerly the Google self-driving car project, Waymo is now looking to integrate Gemini AI chatbot as an in-car assistant.
In a blog post, tech blogger and app researcher Jane Manchun Wong mentioned that while examining Waymo’s mobile app code, she uncovered the complete system prompt for its unreleased Gemini integration (via TechCrunch). “The document, internally titled “Waymo Ride Assistant Meta-Prompt”, is a 1,200+ line specification that defines exactly how the AI assistant is expected to behave inside a Waymo vehicle,” Wong notes.
Gemini could work much more than a simple AI bot in Waymo’s robotaxis
Wong further adds that the feature’s UI hasn’t been released in the public build, meaning there aren’t any screenshots to share what the integration would look like. However, the app researcher mentions that the system prompt makes it clear that Gemini could do much more than just being a simple in-car chatbot. Gemini integration in Waymo robotaxis can openup possibilities to help riders get answers for their queries, manage certain in-cabin features like temperature, lighting, and music, and provide reassurance when needed.
Just to be clear, Gemini, however, won’t be able to control everything. Functions such as volume, seat adjustment, route changes, and window controls remain off-limits. If a passenger requests an unavailable function, Gemini is instructed to respond with aspirational phrases like, “It’s not something I can do yet.”
The prompts instruct the bot to maintain a distinct personality as a friendly and helpful AI companion, separate from the Waymo Driver technology. The idea is “to enhance the rider’s experience by providing useful information and assistance in a safe, reassuring, and unobtrusive manner.” Responses are kept short, typically one to three sentences, and avoid technical jargon to ensure passengers understand instructions easily.
What Gemini could do, and what it couldn’t
Gemini can also greet riders with pre-approved messages personalized using their first name and access contextual data such as the number of trips a passenger has taken with Waymo. While it can answer general knowledge, it can’t perform real-world tasks like ordering food or making reservations. The AI is also explicitly barred from commenting on live driving events, as code references include phrases like, “Your role is not to be a spokesperson for the driving system’s performance, and you must not adopt a defensive or apologetic tone.”
Speaking of Waymo, let’s not forget that earlier this week, a blackout in San Francisco caused Waymo to pause its driverless car service. However, the company has responded, saying that it will be rolling out a software update to improve robotaxi navigation at disabled traffic lights.
With Gemini integration and ongoing updates, Waymo is clearly aiming to make its autonomous ride experience smarter, safer, and more user-friendly, while blending advanced AI assistance with improved real-world navigation.
Source: https://www.timesofai.com/