Incident sheet · /security
- System
- Security & Risk
- Filed
- Read
- 3 min · 623 words
- Severity
- Routine
Google extends Gemini temporary chat retention to two months
Google will retain temporary Gemini chats for around two months, up from 72 hours, while separating chat history from permission to train generative AI models.

Event log
- Temporary Gemini chats will persist for around two months, versus 72 hours previously.
- An archived Gemini Apps Privacy Hub version dated November 12 recorded the 72-hour policy.
- OpenAI temporary chats may remain for up to 30 days; Anthropic incognito chats default to 30 days.
- Google is separating chat-history retention from consent to train generative AI models.
- Both new settings default to off for users who previously disabled Keep Activity.
Daniel Okafor
How long will Google keep temporary Gemini chats?
Google will retain temporary Gemini chats for around two months, up from 72 hours, according to its Gemini Apps Privacy Hub. The policy covers temporary chats and conversations made with chat history turned off. The global Gemini app rollout will take place over the coming weeks.
Google says the change aims to improve safety and security as AI systems become more capable. An archived version of the Gemini Apps Privacy Hub dated November 12 recorded the earlier 72-hour period. The company did not disclose an exact day count for the new window.
For users who selected temporary conversations to limit persistence, the announced shift expands the stated retention period from three days to nearly two months, a substantial change in the data lifecycle.
How does Google's retention compare with OpenAI and Anthropic?
OpenAI says it may keep temporary chats in ChatGPT for up to 30 days. Anthropic retains incognito chats in Claude for 30 days by default. Google's roughly two-month period is about twice either 30-day benchmark.
Google qualifies its figure as “around two months.” The company has not promised a fixed term in the announcement. Its duration exceeds the published 30-day windows of both rivals.
Those comparisons describe stated service policies rather than a universal ceiling on retention, because legal or safety reasons, or investigations into policy violations, may keep some data longer than the headline period.
What will the separate history and training controls do?
Google is splitting chat history from AI-training consent. Until now, both functions sat behind one “Keep Activity” setting. Users can now retain conversations in their chat history without allowing those conversations to train Google's generative AI models.
The new arrangement separates storage from model development. It gives users a direct choice on training while preserving a separate decision about whether their conversations remain in history.
For enterprise buyers, retaining a conversation no longer requires the same approval as training with it, although the announcement does not describe enterprise administration features.
What happens to existing Keep Activity settings?
Both new settings will default to off for users who previously disabled Keep Activity. For users who had Keep Activity enabled, chat history will remain enabled. Those users can separately decide whether Google may use their conversations to train its generative AI models.
Existing activity choices therefore carry over selectively, rather than forcing a single reset of both permissions. The migration does not shorten the new temporary-chat period. Google has not said how users will receive the additional training control.
The transition will not standardize behavior across every account: former Keep Activity opt-outs receive two disabled controls, while existing opt-ins keep history active and retain a separate training decision.
Why is Google extending retention, and what is still unknown?
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 hub says temporary conversations are “not used to train AI models.”
Google did not answer what prompted the policy change. It also did not say whether the longer period affects how conversations can be accessed or reviewed. The company directed questions to its blog post and updated privacy documentation.
The announcement therefore explains the stated safety purpose but leaves operational exposure unresolved, because access, review, and exact expiration mechanics remain unspecified despite the substantial extension beyond 72 hours.
For security teams, the evidence supports a narrow conclusion. Google has disclosed the retention change, global rollout, and split settings. It has not disclosed the policy trigger, access mechanics, or review arrangements for the stored chats.
via techcrunch.com (Source)
More from Daniel Okafor
Show full bio
Staff writer covering technology and product news in enterprise information technology and the leaders who run at CIO Monitor.
203 articles
The newsroom digest
One email a day with the stories that matter. No noise.
Cross-references
- Google Brings Temporary Threads to AI Mode in Search
- Google to lock Gemini Pro behind $19.99 AI Pro plan from October 9
- Google Extends GSA OneGov Gemini Deal Through Feb. 2027
- OpenAI's Businesses in ChatGPT Takes Aim at Salesforce, SAP and Google
- Japan Flags Privacy Risks in Google, Meta, TikTok and LINE Yahoo AI Practices