Google announced separate Gemini controls for keeping chat history and opting out of using conversations to train generative AI models. The announcement also says future chats remain for around two months when the new History setting is off, for purposes including responding to users, processing feedback, and protecting Google, its users, and the public. Google’s announcement was published on November 16, 2026.

How do Gemini’s history and training controls differ?

The announced controls separate two choices: whether to retain chat history and whether to use conversations for model training. Google says users can keep their history while opting out of training. They can also adjust how long history is saved and review or delete it manually.

For a broader explanation of how an assistant’s memory, chat history, and deletion differ, see NeoTeo’s guide to managing memory in ChatGPT, Gemini, and Claude.

Turning off History changes the history setting; Google says future chats are still retained temporarily.

What happens to chats when History is off?

Google says future chats are retained for around two months when the new History setting is turned off. During that period, Google says it uses them to respond to users, process feedback, and support the protection and safety of Google, its users, and the public. Opting out of model training and turning off History are therefore distinct choices, and neither description means that chats are erased immediately.

When did Google plan to roll out the controls?

In its November 16 announcement, Google said a global rollout would begin in the coming weeks.