Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
ReleaseInsight · On RadarTechCrunch→ original

Google will keep Gemini temporary chats for two months instead of 72 hours

Google is extending retention of Gemini temporary chats and conversations with chat history turned off, from 72 hours to roughly two months. The company cites safety and service protection, but did not answer questions about why it made the change. OpenAI keeps temporary chats for up to 30 days, and Anthropic keeps incognito chats for 30 days by default. So a "temporary chat" in Gemini no longer means "deleted almost immediately."

AI-processed from TechCrunch; edited by Hamidun News
Google will keep Gemini temporary chats for two months instead of 72 hours
Source: TechCrunch. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTechCrunch ↗
Editor-in-Chief
Key metric #1Google says temporary chats are not used to train models
Key metric #2Chat history and AI training are now two separate settings
Key metric #3Users who had turned off Keep Activity will see both new options switched off

Many people assume a temporary chat with an AI is like a conversation in a kitchen: you talk, then you forget. Google has just reminded everyone that on the servers, things work differently. The company is increasing the retention period for temporary chats in Gemini, and for conversations held with chat history turned off, from 72 hours to roughly two months. The change will roll out worldwide in the Gemini app over the coming weeks.

The previous 72-hour period was stated in the Gemini Apps Privacy Hub. This is shown in an archived version of the page dated November 12, which TechCrunch cites. The period has now grown almost 20-fold.

How this compares with competitors

The main rivals have shorter periods. OpenAI says it may keep temporary chats in ChatGPT for up to 30 days. Anthropic keeps Claude incognito chats for 30 days by default. That means Google now keeps such conversations roughly twice as long.

A caveat applies to all three companies: data may be kept longer if legal reasons, security concerns, or investigations into violations of service rules require it. Retention periods in policies are usually the lower bound of what actually happens to data.

Google did not explain why it is making the change now. It did not respond to TechCrunch's questions about its motives or about whether the way conversations are accessed and reviewed will change. Instead, it pointed to a blog post. That post says chats are kept longer to respond to users, process feedback, and protect services and people. The updated Privacy Hub also cites possible system failures as a reason. The company's broader argument is that AI is becoming more powerful, so more safety measures are needed.

Model training and history are now configured separately

At the same time, Google is changing its settings. Previously, both chat history and the use of conversations for model training were tied to a single Keep Activity switch. If you wanted to see your old chats, you had to agree that they would be used for training. Now these are two independent parameters. You can keep history stored while not letting your conversations be used to train Google's generative models.

For users who previously turned off Keep Activity, both new settings will be off by default. Users who kept Keep Activity on will retain their chat history. They can decide separately whether their conversations are used for training. Google also stresses that temporary chats are not used for model training.

What this means in practice

For an ordinary user, the takeaway is simple. If you previously turned off history for privacy reasons, keep in mind that your conversations still sit on Google's servers for about two months. Secrets, customers' personal data, and work documents in such a chat remain on the company's servers for that entire time.

For developers and product managers, the lesson is broader:

  • If your company allows Gemini use, it is worth rereading internal rules for working with AI. They often say "use a temporary chat" as if that guarantees quick deletion.
  • Sensitive data calls for enterprise plans with contractual guarantees, not the consumer app. These terms are not affected by this news, but they are worth checking separately.
  • If you are building your own AI product, conversation retention periods and separate consent for training are becoming a competitive argument. Users compare privacy policies, and the major players have just shown how much they differ.

There are also honest limitations. Google's statements do not make clear who can read saved chats or under what conditions. It has only said that temporary conversations do not go into training. Whether to change settings and how to think about "temporary" modes is still up to users. But the assumption that "everything disappears in three days" no longer holds for Gemini.

Story

Key Questions Answered

How this compares with competitors?

The main rivals have shorter periods. OpenAI says it may keep temporary chats in ChatGPT for up to 30 days. Anthropic keeps Claude incognito chats for 30 days by default. That means Google now keeps such conversations roughly twice as long. A caveat applies to all three companies: data may be kept longer if legal reasons, security concerns, or investigations into violations of service rules require it. Retention periods in policies are usually the lower bound of what actually happens to data.

What are the key details on Model training and history are now configured separately?

At the same time, Google is changing its settings. Previously, both chat history and the use of conversations for model training were tied to a single Keep Activity switch. If you wanted to see your old chats, you had to agree that they would be used for training. Now these are two independent parameters. You can keep history stored while not letting your conversations be used to train Google's generative models. For users who previously turned off Keep Activity, both new settings will be off by default.

What this means in practice?

For an ordinary user, the takeaway is simple. If you previously turned off history for privacy reasons, keep in mind that your conversations still sit on Google's servers for about two months. Secrets, customers' personal data, and work documents in such a chat remain on the company's servers for that entire time. For developers and product managers, the lesson is broader: If your company allows Gemini use, it is worth rereading internal rules for working with AI.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…