Google brings Gemini’s personal-data lookups to Android Auto

Google began rolling out Personal Intelligence for Gemini in Android Auto on October 20, 2026, letting the in-car assistant draw on connected account information rather than requiring drivers to explain every detail. The feature is optional, but people who already enabled it in Gemini will bring those permissions into the car.

Android Auto gets the upgrade first. Google says vehicles with Google built-in will receive it in the coming months. The distinction is where the software runs: Android Auto projects a phone-based interface onto the vehicle’s display, while Google built-in integrates Google software and services directly into the car. The announcement covers both routes, but their Personal Intelligence rollouts are not simultaneous.

From an address to “my appointment”

Google describes two jobs for Personal Intelligence: finding a specific detail and reasoning across multiple sources. With permission, it can connect to Workspace, including Gmail and Google Calendar, as well as Google Photos, Maps, YouTube and Search. The intended change is that a spoken request can refer to information already stored in those services, instead of making the driver supply the background.

That turns an ordinary navigation request into a personal lookup. Google’s example is asking for directions to a doctor’s appointment without remembering the street address. It also says Gemini can locate a car wash from a photo taken the previous week and start navigation there. These are company-described uses of the connected information, not independently verified results.

Other requests Google says it can handle

• Flight details: Ask for a flight’s gate and whether it is on time. Gemini is designed to combine the booking information with a delay lookup and answer both questions together.
• Restaurant preferences: Ask for an indoor restaurant along the route that fits familiar tastes. Google says a vegetarian user could receive a shortlist suited to that preference.
• Conversation continuity: Refer to the coffee shop discussed the day before, rather than naming it again. Google says Gemini can retrieve that context and provide directions.

An earlier opt-in follows the driver

The feature is off by default, and users can choose which apps to connect or turn it off later. But Google is not describing a fresh permission decision for every setting where Gemini appears. If Personal Intelligence is already enabled in the Gemini app, the company says it comes into the car with the user.

Ars Technica notes that other people may interact with the screen in a car, making the existing account choices worth reviewing. Its practical guidance is to open the Gemini app’s settings and find Connected Apps. That section contains controls for services such as Workspace, Search and Photos, letting users reconsider which sources should inform responses.

For the later Google built-in rollout, Google specifies an additional condition: personal responses are shared only when the user is in the car with their phone connected. That is a stated access safeguard for the integrated vehicle system. Google’s announcement directs users to Gemini settings to manage their preferences.

Permission controls also differ from data-training rules. Ars reports that Google says it does not train Gemini on personal data itself, but does use users’ AI inputs and outputs. Those outputs can contain personal information drawn from connected services. According to Ars, Google attempts to filter that information out but does not guarantee its removal. Choosing connected apps therefore governs what can inform an answer; it should not be confused with a blanket promise about every response generated from that material.