Google is adding separate controls for AI training and chat history in the Gemini app, according to the source. Users will see distinct toggles for model training and chat history in the Gemini activity settings, and the change is rolling out globally, though it may take a few weeks to appear for every account.

Previously, the only way to stop Google from using chat contents, including attachments, for training was to turn off chat history. That also made it impossible to return to earlier conversations. Under the new setup, AI training is opt-out. Users can find the training toggle under Activity in the Gemini settings, or through their Google account at Data and Privacy, then My Activity, then Other activity, then Gemini Apps Activity.

The change applies only to future interactions. Data already submitted, including inputs, responses, uploaded files, tool calls, location, and data from connected apps, remains part of the training data. The source says Google may retain this data for up to three years to improve its models. Google has said it tries to filter personal information from Gemini activity history, but the source notes this is not a guarantee. Reviewed chats already used for training will not be deleted retroactively.

Opting out of training does not stop Google from holding the data. The source says Google may use it in other ways for up to three years, just not as AI training data.

For users who want no retained data, the options are narrower. Currently, if history is disabled, Google keeps that data for 72 hours. Once the updated page is live, the retention period will be two months. The source quotes Google saying this is to "respond to you, process your feedback, and protect Google, its users, and the public."