GM bets on a deeply integrated in-vehicle AI system that leverages GM data and OnStar to go beyond Google's Gemini

General Motors plans to launch a proprietary in-vehicle AI system later this year that goes beyond Google's Gemini by tapping into vehicle telematics, OnStar intelligence, and proprietary GM data. For AI builders, this signals a shift from general-purpose assistants to deeply integrated, vehicle-specific AI that could redefine how automakers deploy conversational AI.

What happened

GM announced it will launch a more deeply integrated native AI assistant later this year, combining conversational AI with GM vehicle knowledge and OnStar intelligence. The system is designed to understand the vehicle, the drive, and customer needs to simplify ownership, according to Anna Santos, GM's director of product management of voice and AI/machine learning.

GM is working with an unnamed large language model provider to support predictive maintenance, vehicle telemetry, and auto-focused features. One example is a "kids setting" command that tailors music, seats, heating/cooling, and door locks for children. The data powering these features will be proprietary to GM.

This follows GM's earlier rollout of Google's Gemini AI assistant in millions of 2022 model-year and newer vehicles. Gemini offers natural language conversation, live sessions, and basic climate/radio control, but Santos described it as operating like a phone assistant with limited vehicle integration.

Why AI builders should care

GM's move signals that automakers want to own the AI layer by combining external LLMs with proprietary vehicle data. This creates a new pattern: OEMs partner with LLM providers but keep the data and integration in-house. For AI builders, this means opportunities to build middleware that securely accesses vehicle telematics and OnStar data for specialized tasks.

The system's focus on context-specific commands (like "kids setting") shows how vehicle-specific AI can deliver experiences that general-purpose assistants cannot. This also raises questions about data privacy and control within the OEM ecosystem, which developers will need to navigate when building for automotive platforms.

Practical implications

For product teams and developers, GM's approach suggests several practical takeaways:

• Middleware opportunities: There is a need for interfaces that securely bridge vehicle telemetry data with LLM APIs while maintaining OEM data governance.
• Customization patterns: The "kids setting" example illustrates how automakers can use AI to combine multiple vehicle subsystems (HVAC, audio, locks) into single voice commands. Developers building for automotive should expect more such compound commands.
• Data privacy boundaries: GM's proprietary data strategy means third-party developers may have limited access to raw vehicle data. Building on top of OEM-approved APIs will be essential.

Caveats

Several details remain undisclosed. GM has not named its LLM provider, shared specific deployment timelines beyond "later this year," or released technical specifications. The feature set described (predictive maintenance, telemetry, kids setting) may evolve before launch. Much of the information comes from business press announcements and GM statements, not a fully disclosed technical spec. AI builders should watch for more concrete details closer to launch.