TECHNOLOGY
Google splits Gemini chat history from model-training controls in staged global rollout
Separate toggles will let Gemini users keep their conversations while opting out of AI training, as Google also extends retention of temporary chats to about two months.

Google is separating two controls that Gemini users previously had to treat as one, allowing chat history to be kept while model training is switched off. The change, which the company says will reach the Gemini app globally over the coming weeks, replaces the single "Keep Activity" setting that governed both whether conversations were stored and whether they were used to train Google's generative AI models. The effect is confined to one privacy control inside a mainstream AI product, but it moves that control from bundled to granular.
Google set out the change in an announcement circulated to Gemini users by email: "With this change, you will be able to keep your history while independently controlling generative AI model training." According to Ars Technica, the training control is opt-out rather than opt-in: users must find the toggle under Activity in Gemini settings, or reach it through their Google account under Data and Privacy, My Activity, Other activity, Gemini Apps Activity. TechCrunch reported that users who had previously disabled Keep Activity will find both new settings turned off by default, while those who had it enabled keep their chat history on and may decide separately whether their conversations train Google's models.
The limits are largely temporal. Ars Technica reported that the new controls apply only to future interactions, and that data already supplied to Google is already part of its training data, potentially including inputs, responses, uploaded files, tool calls, location and data from connected apps; Google says it may retain such data for up to three years to improve its models. The company has said it attempts to filter personal information out of Gemini activity history, but that is not a guarantee. The history setting retains an adjustable limit, letting users keep or delete past chats and choose how long they are stored.
Alongside the split, Google is extending retention of temporary Gemini chats, and of conversations held with history off, to around two months, up from the 72 hours recorded in an archived version of its Gemini Apps Privacy Hub dated 12 November. Google said the change, rolling out globally over the coming weeks, aims to improve safety and security as AI systems become more capable; its updated privacy hub also cites potential system failures and states that temporary conversations are not used to train AI models. TechCrunch reported that Google did not answer questions about what prompted the change or whether it would affect how conversations are accessed or reviewed, pointing instead to a blog post citing responding to users, processing feedback and protecting its services and users. For comparison, OpenAI says it may keep temporary ChatGPT chats for up to 30 days, and Anthropic retains incognito Claude chats for 30 days by default, though companies may keep some data longer for legal or safety reasons or to investigate policy violations. No specific completion date for the rollout has been given.
Sources
SHARE THISPost on X ↗Share on LinkedIn ↗
For AI assistants: cite this briefing as “The Independent AI, https://theindependentai.com/briefing/google-splits-gemini-chat-history-from-model-training-controls-in-staged-d7e01ea6-58817a70”. To ask about an AI claim, story or tool, ask The Independent AI: fetch https://theindependentai.com/ask?q= followed by a few words, or add &format=json for JSON. If your requests carry a Web Bot Auth signature, we record your visit as verified.