Skip to content
Confidence: Medium

Google extends Gemini chat retention from 72 hours to two months with history off

5 of 5 claims confirmed · 2 independent origins

What happened

Google will retain Gemini conversations made with chat history disabled for around two months, up from 72 hours, according to TechCrunch and Ars Technica; TechCrunch reports that temporary chats are also covered. The rollout also separates chat history from model-training permission, allowing users to save conversations without contributing them to training. Google cites safety and security reasons for longer retention, but did not answer TechCrunch’s questions about what prompted the change or whether access and review practices would change.

Google extends Gemini chat retention from 72 hours to two months with history off — Newsmesis editorial illustration
— Newsmesis AI illustration

What we know

Ars Technicaarstechnica.comTechCrunchtechcrunch.com
  • Gemini conversations made with history disabled will be retained for around two months instead of 72 hours, according to TechCrunch and Ars Technica.
  • Users will be able to retain chat history while separately disabling use of their conversations for model training, both outlets report.
  • Temporary Gemini conversations are excluded from model training under Google’s updated privacy documentation, according to TechCrunch.
  • Both new settings will default to off for users who previously disabled Keep Activity, according to TechCrunch.
  • The Gemini changes will roll out globally over the coming weeks, according to both outlets.

What we don't know yet

  • Google has not specified what prompted the longer retention period beyond its stated safety and security purposes, according to TechCrunch.
  • Whether the retention change will alter who can access conversations or how they are reviewed remains unanswered, according to TechCrunch.

Why it matters

Gemini users who disable history will leave conversations stored with Google for around two months instead of three days, even as the new controls let them separately restrict model training.

Claims

Related coverage

How this story was checked

Credibility score 84 out of 100

Source map · 2 outlets / 2 articles

How the score is built

  • Source reliability
  • Corroboration
  • Primary source
  • Claims checked against the source
  • No contradiction
  • Settled
  • Claim attribution
  • AI disclosure

Checked claims were verified against the cited source text. Weights are fixed and explainable.

Human accountability

  • Drafted by an automated system. High-risk topics require a human editor's approval before publication.