General Motors is rebuilding OnStar around artificial intelligence, according to Autoguide's report from the GM Forward event in Manhattan. The changes start with the 2027 Cadillac LYRIQ and include a new OnStar AI Assistant and a redesigned mobile app, both due in 2027.
What the assistant does
The OnStar AI Assistant draws on real-time vehicle diagnostics, contextual driver information and official GM vehicle knowledge. Drivers wake it with the phrase "Hey OnStar." Autoguide lists commands such as adjusting cabin temperature, opening the EV charge port, and checking tire pressure and EV range.
It can also explain dashboard warning lights, tailored to the specific vehicle. If it cannot answer a question, it routes the driver to a live human advisor. The assistant is conversational, in the style of other large language model agents.
Autoguide notes the assistant is not exclusive to new hardware. It will reach the LYRIQ through an over-the-air software update, with other GM models to follow. The report does not say which models or when.
One app, connected cameras
GM will merge the existing OnStar Guardian app and the individual Buick, Cadillac, Chevrolet and GMC apps into a single OnStar app, scheduled for 2027.
The app adds Connected Cameras, which record automatically when triggered by vehicle events such as horn activation. Clips are uploaded securely to the app and the driver is alerted. It also adds OnStar Family, with location and destination sharing, driver oversight tools and community safety notifications, and Predictive Safety Alerts built from vehicle sensor data, local weather, road parameters and navigation routes.
Privacy controls
GM said the AI and camera features are opt-in. Users can switch camera, microphone, location and data-sharing permissions on or off. Video recordings carry time limits and end-to-end encryption, and GM said it will not have access to private Connected Camera footage.
Fleet products
For commercial customers, GM is introducing OnStar Fleet Intelligence and an OnStar Fleet AI Assistant. According to Autoguide, fleet managers can query vehicle data in natural language to monitor vehicle health, optimize energy and fuel consumption, schedule maintenance and track asset utilization across operations.
The source does not give pricing, availability dates for the fleet products, supported vehicle models, or details on integration with third-party fleet systems.
