Google extends Gemini history-off chat retention to two months
Google will retain history-off Gemini chats for around two months, up from 72 hours, and split chat-history and AI-training controls during a global rollout.

- Google will retain temporary Gemini chats for around two months, up from 72 hours.
- The global Gemini app rollout is scheduled for the coming weeks; the prior policy appears in a Privacy Hub archive dated November 12.
- OpenAI says temporary ChatGPT chats may be kept for up to 30 days; Anthropic says Claude incognito chats default to 30 days.
- Google is replacing the combined Keep Activity control with separate chat-history and AI-training settings.
- Previous Keep Activity opt-outs default both new settings off; previous opt-ins keep chat history on and receive a separate training choice.
By Tom Whitfield
Google will keep temporary Gemini chats for around two months, up from 72 hours, and roll the change out globally in the Gemini app over the coming weeks.
That extension changes the practical meaning of “history off.” Temporary chats (Gemini conversations created with chat history turned off) fall under the new window even though users have disabled chat history. The announced period is roughly two months rather than three days.
The previous 72-hour limit appeared in an archived version of Google’s Gemini Apps Privacy Hub dated November 12. Google describes the change as improving safety and security as AI systems become more capable. The company says the update will roll out globally over the coming weeks.
Google says it retains the chats longer to “respond to users, process feedback, and protect its services and users.” Its updated Gemini Privacy Hub also cites potential system failures as a reason for retention. The company says temporary conversations are “not used to train AI models.”
Google’s announcement establishes the longer policy, but leaves two operational questions open. Google did not answer what prompted the revision or whether the longer period would affect how conversations are accessed or reviewed. It pointed instead to a blog post for its stated purposes of retention.
By comparison, OpenAI says it may keep temporary chats in ChatGPT for up to 30 days, while Anthropic retains incognito chats in Claude for 30 days by default. On the periods described, Google’s roughly two-month window is longer. The comparison has limits: companies may keep some data longer for legal or safety reasons or to investigate policy violations.
The policy language also differs. OpenAI uses “up to,” Anthropic describes 30 days as the default for incognito chats, and Google says “around two months.” The cited retention figures therefore do not amount to an unconditional deletion deadline.
A split between history and training
Google is separating two decisions that previously arrived as one. A single “Keep Activity” setting governed both whether Gemini kept conversations in chat history and whether Google could use them to train its generative AI models. Users can now retain their history while refusing training use.
Google will apply the new settings according to the user’s previous position:
- Keep Activity off: Both the chat-history and AI-training settings will default to off.
- Keep Activity on: Chat history will remain enabled, while training becomes a separate decision.
What does this change for Gemini users?
Users who deliberately disabled Keep Activity will have both new settings off by default, but they face a material change in how long Google says it keeps temporary exchanges. Users who enabled Keep Activity keep their chat history and gain a separate decision over training.
For users who want a conversation absent from history and outside model training, both choices can remain off. That configuration still does not create a 72-hour deletion window. Google’s explanation for the longer period concerns service response, feedback, protection and system failures, not a new training policy for temporary chats.
The practical change is separation of functions. History controls whether Gemini displays prior conversations; training controls whether Google can use conversations to improve its generative AI models. Longer storage for temporary chats remains a separate policy. Users with history enabled can keep a conversation for reference without allowing that conversation to become training material.
Monitor the global rollout over the coming weeks, the final Privacy Hub language and whether Google clarifies how retained temporary chats can be accessed or reviewed.
via techcrunch.com (Source)
More from Tom Whitfield
Show full bio
News editor covering regulation, policy and their impact on Search rankings and algorithm updates, generative engine optimization at SERP Journal.
183 articles