GM to launch its own in-vehicle AI assistant later this year

Carmakers rolling out AI assistants is not new, but the sequencing here is what makes GM interesting. Only months after pushing Google's Gemini out to roughly four million vehicles on the road, the company is telling reporters it is building its own on top. CNBC reports that GM plans to launch a proprietary in-vehicle AI assistant later this year, one that goes beyond what the Google-powered system can do.

The person making the pitch is Anna Santos, GM's director of product management for voice and AI/machine learning. Her framing, as reported, is that the new assistant will combine conversational AI with GM vehicle knowledge and OnStar intelligence to create capabilities a general-purpose assistant cannot. That is the tell. GM is not trying to out-model Google on language. It is trying to own the layer that sits between the driver and the parts of the car only GM sees: telematics, service history, the OnStar data pipe it has been building for a quarter century.

The strategic logic writes itself. If Gemini is the mouth and ears, GM wants to keep the brain that knows the vehicle. Predictive maintenance, vehicle telemetry, and personalization touches like a reported 'kids setting' are the kinds of features that turn an assistant from a novelty into a reason to renew an OnStar subscription. It also keeps a large chunk of driver data, and any future in-car commerce that rides on it, inside GM's own stack rather than a partner's.

The honest caveat is that the reporting is thin on specifics. GM has said it is working with an unnamed large language model provider, which means the dependency question has not gone away, only moved. It is not yet clear whether the native assistant replaces Gemini in the cabin or runs alongside it, what the launch window inside 'later this year' actually is, or which model years qualify. Take the specifics as reported, not settled.

What is worth watching is whether other automakers follow the same two-step: license a frontier assistant to catch up on conversational quality, then quietly build the vehicle-aware layer that keeps the customer relationship yours.