Google first put Gemini into Android Auto last year, then expanded the assistant to cars with Google built-in in April. Now it is taking the next step: Personal Intelligence, which lets Gemini draw on information held across Gmail, Calendar, Maps, Photos, YouTube and Search.¹

The pitch is straightforward. A driver heading to the airport could ask for a flight status, with Gemini finding a confirmation email in Gmail before checking the journey. Someone could say, “take me to my doctor’s appointment,” or ask the assistant to locate a car wash pictured last week and start navigation.¹ Google says the purpose is to reduce the detailed back-and-forth that can distract from driving.

The company stresses that the system is opt-in. Personal Intelligence is rolling out to Android Auto now and is due in Google built-in vehicles in the coming months; without it, Google says drivers can still navigate, play music, make calls and ask general questions, just without tailored answers.¹

But that opt-in is where the unease begins. One account of the rollout notes that Personal Intelligence began as a Gemini feature outside the car and may already be enabled by users who accepted Google’s prompts elsewhere. Those connected-app settings can carry over to Android Auto, meaning a driver may encounter the new capability without freshly considering what Gmail, Calendar, Photos and Search access entails.²

There are further limits intended to contain exposure: in Google built-in cars, personal responses require a phone connection over Bluetooth, and in-car settings can disable the feature. Yet the privacy concern is not simply whether Gemini trains on personal data—Google says it does not—but that interactions and outputs may still contain sensitive details, which the company says it tries, but cannot guarantee, to filter out.²

For drivers, the upgrade offers a sharper digital co-pilot. It also makes a mundane settings check newly consequential: the car may be hands-free, but the personal data behind its answers is not.