GM is developing its own in-vehicle AI assistant since Google’s Gemini is unable to access the information that the car possesses.

      TL;DRGM plans to introduce a proprietary AI assistant later this year that will surpass Gemini by incorporating OnStar, vehicle telemetry, and predictive maintenance. The provider of the large language model remains unnamed. Gemini will still be available.

      General Motors is set to launch an in-car AI assistant later this year, designed to tap into vehicle data that Google’s Gemini cannot access. This new system will work with OnStar, GM’s telemetry systems, and exclusive vehicle information to offer predictive maintenance, real-time diagnostics, and vehicle-specific commands that a general-purpose assistant cannot deliver. GM is collaborating with an undisclosed large language model provider for this technology.

      “There’s a limit to what an AI that’s just sort of sitting at the top level of the vehicle can do,” remarked Anna Santos, GM’s director of product management for voice and AI. Gemini was introduced for eligible 2022 and newer GM vehicles earlier this year to manage conversational requests, temperature, and radio settings. The new GM assistant will have a deeper comprehension of the vehicle's mechanical condition, driving habits, and owner preferences. Santos illustrated this with a “kids setting” command that would adjust music, seating, climate, and door locks all at once.

      This differentiation is significant from a commercial standpoint. Gemini operates like a phone assistant displayed on the dashboard, while GM’s proprietary system utilizes data owned by GM, not Google. “It’s data that’s going to be proprietary to GM,” Santos noted. Google has been integrating AI into all its managed products, and GM’s decision to develop a parallel system instead of relying solely on Gemini stems from concerns similar to those raised by Nadella about companies placing their decision-making in the hands of a single AI provider.

      GM is not eliminating Gemini; both assistants will operate simultaneously. The Gemini layer will manage general questions and entertainment, while the GM layer will handle all information the car holds about itself. GM is also advancing its autonomous driving capabilities towards a robotaxi-ready feature, and an in-car AI that grasps predictive maintenance and telemetry will facilitate fleet management for autonomous operations. Santos referred to it as “the beginning of a broader AI journey.” The unnamed LLM provider indicates that GM is keeping its options flexible regarding which model powers the system, thus avoiding becoming dependent on a single vendor as a result of its partnership with Gemini.