Google brings Gemini to cars with Google built-in, will update millions of vehicles

Google has started rolling out Gemini in cars with Google built-in, replacing the current Google Assistant with a more conversational AI interface. The launch begins in the US in English and covers not only new models but also compatible cars already on the road, which will receive the update over the air.

How the rollout will work

Google announced the launch on April 30, 2026. It quickly became clear this is no niche experiment: the day before, on April 29, General Motors said Gemini would arrive in about 4 million vehicles from model year 2022 onward across Cadillac, Chevrolet, Buick and GMC. Google did not limit the announcement to GM, so the rollout will likely extend to other automakers that already use the Google built-in platform.

For drivers, the switch will look like an ordinary system update. If the car is compatible and the owner is signed in to their Google account, the screen will offer a switch from Google Assistant to Gemini. Once enabled, the assistant can be summoned by voice, with a steering-wheel button or via a microphone on the home screen. Google stresses that the feature will first appear in the US and only in English, then expand to other countries and languages in the coming months.

What it can do on the road

The main change is the move from rigid voice commands to a more natural dialogue. Instead of short instructions, drivers can speak almost as they would to a person: for example, ask for a good restaurant along the route with indoor or outdoor seating, then follow up on whether parking is nearby and whether the dishes fit specific dietary restrictions. Gemini uses Google Maps data and holds context better, so a chain of follow-up questions does not break off after the first reply.

These are the scenarios Google and GM show as core to the new version of the assistant:

• finding stops along the route, refined by rating, parking and menu
• summarizing new messages and sending a reply without a phone, including last-second edits
• controlling music, radio and streaming with looser requests rather than exact station names
• traffic tips, events along the way and reports of accidents or other obstacles on the road
• Gemini Live conversation mode for learning, brainstorming and questions about the destination

Google is separately promoting Gemini Live, which is still in beta. It can be turned on with a button in the interface or with the command "Hey Google, let's talk." In this mode the assistant suits not only utilitarian tasks like routing or cabin climate but also longer conversations: you can ask for a short history of the place you are heading to, help planning a trip, a discussion of ideas before a meeting, or even a rehearsal of a difficult conversation.

Deeper into the car's systems

The most interesting part of this launch is not only the conversational interface but also the tighter integration with the car itself. Google says Gemini will be able to answer questions using data from the owner's manual provided by the automaker. Queries such as "how do I prepare the car for an automatic car wash" or "how do I limit the trunk opening height in a low garage" should return an answer tied to the specific model. However, the depth of such answers will depend on the brand and trim level.

Electric vehicles get another useful layer. Drivers can ask about the current battery charge, the predicted charge on arrival and nearby chargers, then immediately ask whether there is a cafe or another place to spend time while charging. Control of familiar car functions in natural language also remains: if you say the cabin is cold and the windshield is fogging up, Gemini should understand the intent and turn on the heating and defog by itself.

Google has also already named the next stage: the assistant will get a deeper link to Gmail, Google Calendar and Google Home. This means the car is gradually turning from a standalone screen with navigation and music into one more access point to Google's personal ecosystem, with calendar, email, home devices and trip context in a single interface.

What this means

Google is betting that generative AI will become not a separate app but a new layer for controlling the devices around the user, including the car. For automakers, it is a way to quickly add a smarter interface to the cabin without large in-house AI development. For drivers, it is a chance to get a genuinely useful hands-free tool, if it proves accurate, unobtrusive and safe while driving.