What happened
Google is rolling out separate settings in the Gemini app that let users decide independently whether their chats are saved as history and whether they are used to train its AI models, Ars Technica reported. Until now, people who wanted to keep their conversations out of training had to switch off chat history entirely, which meant they could not return to an earlier conversation later.
Training stays on by default, so users must find the new toggle themselves. It sits under Activity in the Gemini settings and can also be reached through the Gemini Apps Activity page in the Data and Privacy section of a Google account. The change covers only future use. Material already submitted, such as prompts, responses, uploaded files, tool calls, location and data from connected apps, remains in the training data, and reviewed chats are not removed retroactively.
Turning off history will also mean Google keeps those chats for two months, up from 72 hours today. Google said the longer window lets it respond to users, handle their feedback and protect itself, its users and the public. The company said the update is reaching users worldwide but may take a few weeks to appear for everyone.
Context
Gemini is built into many Google products, and greater use gives Google more data to train and improve its chatbot, while ways to limit that had been scarce. Google says it may keep Gemini data for up to three years, and opting out of training does not stop it from using that data in other ways. Google has previously noted that it tries to strip personal details from Gemini activity history, though that filtering is not guaranteed.
The briefing
The week in tech, in one email.
The stories worth your time, every week. Unsubscribe anytime.
