Skip to content
Marketing, IT & AI automation · Surrey and Metro Vancouver
Digital Allo Marketing Services Ltd. Contact Us

Digital Allo Blog · November 19, 2026

Your “Temporary” Gemini Chats Are Now Kept for Two Months — What Surrey Businesses Should Do

By Digital Allo Editorial Team Published

“Temporary” used to mean three days. Now it means two months.

If anyone on your team uses Gemini — to draft a quote, summarise a client email, reword a proposal, or check a price — there is a change you should know about. On 16 November 2026, Google updated its Gemini privacy controls. Buried in that update: temporary Gemini chats, and chats made with History turned off, are now kept for around two months. Until last week, the stated period was 72 hours.

TechCrunch spotted the change on 17 November by comparing Google’s Gemini Apps Privacy Hub with an archived copy from 12 November. The old page said 72 hours. The new page says “around two months” — for both temporary chats and chats you have when your History setting is off. That is roughly twenty times longer, and Google’s announcement did not mention the old figure at all.

The change is rolling out globally in the Gemini app over the coming weeks, so the setting may not have reached every phone yet. But the direction is set: the chat you thought vanished is now sitting on Google’s servers for about two months.

What actually changed — and what didn’t

What changed: the retention window. A temporary chat — the kind you start precisely because you don’t want it kept — used to be held for 72 hours. Now it is held for around two months, tied to your account, even though it never appears in your history. The same applies to any chat made with History switched off.

What also changed: the settings themselves. Until now, Gemini’s chat history and AI training were tied to a single “Keep Activity” switch. Google has split them into two separate settings. You can now keep your conversations in your history without letting Google use them to train its generative AI models. Both new settings will be turned off by default for people who previously disabled Keep Activity; people who had it on keep their history on, and can now decide separately about training.

What didn’t change: Google says temporary chats are not used to train its AI models. That statement is in the updated Privacy Hub, and it is worth knowing — but it is not the whole story. Google also says chats are kept longer to respond to users, process feedback, and protect its services and users, and the Privacy Hub cites potential system failures as a reason. Google declined to answer TechCrunch’s questions about what prompted the change, or whether it affects how conversations are accessed or reviewed.

One more detail from NotebookCheck’s reading of the updated Privacy Hub: if you send feedback on a Gemini response while History is off, Google collects the last 24 hours of your chats, including attachments. And chats that get reviewed by Google’s service providers can be kept for up to three years, separated from your account, even if you delete them later. So “delete” and “gone” are not the same thing either.

How the others compare

Google is now the outlier on temporary-chat retention. 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. Both companies also note they may keep some data longer for legal or safety reasons, or to investigate policy violations — so nobody’s “temporary” means instant deletion. But Google’s new window, at around two months, is roughly double the others.

That matters less as a ranking of chatbots and more as a habit check. If your team treats “temporary chat” as a safe place for sensitive material, the safety assumption just got weaker — on the chatbot most of your staff probably already have on their phones.

Why a Surrey business should care

This is not about whether you like Google. It is about what your team types into a chatbot during a normal working day. A few examples that happen in Surrey offices every week:

  • A clinic receptionist pastes a patient’s email into Gemini to draft a reply — with the patient’s name and appointment details in it.
  • A contractor pastes a client’s address and a quote into a chatbot to reword it.
  • An immigration consultant pastes a client’s case notes in to summarise them.
  • A bookkeeper pastes a spreadsheet with client names and amounts to “make sense of this”.

None of that is unusual, and none of it is malicious. But each one is a disclosure of someone else’s personal information to a third party — and under BC’s Personal Information Protection Act (PIPA), a business is responsible for the personal information it collects, uses and discloses. This blog is not legal advice, and the right move is to check your own obligations with your privacy advisor. But the practical point is simple: if you would not email that information to a stranger, don’t paste it into a chatbot either — because for around two months, it is sitting on someone else’s server, tied to the account that sent it.

The other risk is quieter: habit. Staff who learned that “temporary chat” means “gone in three days” will keep acting on that belief until someone tells them it changed. Google announced this in a blog post about giving users more control. The retention change was not in the headline.

The 5-step fix: a safe-AI-use rule for your business

Step 1 — Check the new Gemini settings today. Open the Gemini app, go to Settings, and look at the two separate controls: History and AI training. If you previously had Keep Activity off, both should be off by default — confirm it anyway. If it was on, your history stays on and you now need to decide about training separately. Do this on every device your team uses, and on both personal and work accounts — the setting follows the account, not the phone.

Step 2 — Treat “temporary” as “kept for two months”. Tell your team plainly: a temporary chat is not a shredder. It means the chat won’t show up in your history. It does not mean Google deletes it. The same applies to History-off chats. If a piece of information should not sit on Google’s servers for two months, it should not go into the chat at all.

Step 3 — Make a short never-paste list. Write it down, one page, and stick it where your team can see it. Ours would start with: client names with contact details, addresses, health or financial information, quotes that include personal details, passwords and login details, internal pricing or margins, and anything covered by a confidentiality agreement. If the task needs that information, do it in your own systems — not in a public chatbot.

Step 4 — Give your team a proper AI tool instead of a personal account. The reason staff paste client details into Gemini is usually that it is the fastest tool to hand. A custom AI solution or an AI chatbot built for your business can do the same drafting and summarising inside your own systems, with your own data rules — so nobody has to choose between getting the work done and protecting client information. That is the same reason to keep work off personal accounts: a personal Gemini account follows the person, not the business.

Step 5 — Put it in writing and check it once a quarter. A one-page AI-use rule, added to your onboarding and reviewed when the tools change, is worth more than a one-time warning. These settings change without a headline — Google has changed Gemini’s retention, its settings layout and its training controls in the same week. Put a quarterly reminder in the calendar to re-check the privacy settings on the AI tools your team actually uses.

What not to do

Don’t panic and ban AI tools — your team will just use them on personal phones instead, where you have even less visibility. Don’t assume ChatGPT or Claude are automatically safer: their temporary chats are kept up to 30 days, and the same never-paste list applies to them too. Don’t rely on “delete” to fix a paste that already happened — deleted chats can still be held for review, and the retention clock is Google’s, not yours. And don’t leave this to one person: the risk is not one careless paste, it is a whole team acting on an old assumption about what “temporary” means.

The bottom line

Google gave Gemini users a genuine improvement this week — history and AI training are now separate choices, and you can keep one without the other. But it also quietly doubled down on the other side: the chats you thought were temporary are now kept for around two months.

For a Surrey business, the fix is not technical. It is a five-minute settings check, a one-page list of what never goes into a chatbot, and a proper AI tool for the work that needs client information. Do that now, and the next privacy change — because there will be one — is a settings update, not a scramble.

Want Digital Allo to set this up for you? We build custom AI solutions and AI chatbots that work inside your business’s own data rules, and our cybersecurity services and managed IT services cover the settings, accounts and policies around them. Start with a free AI Opportunity Assessment, or talk to us about IT services in Surrey.

FAQ

How long does Google keep Gemini temporary chats? Around two months, as of Google’s change announced 16 November 2026 — up from 72 hours. That covers temporary chats and chats made with your History setting turned off. The change is rolling out globally in the Gemini app over the coming weeks.

Does turning History off in Gemini delete my chats? No. Turning History off stops chats appearing in your history, but Google keeps them tied to your account for around two months. Temporary chats work the same way: they never show up in your history, but they are retained for around two months.

Are temporary Gemini chats used to train AI models? Google says no — temporary chats are not used to train its AI models. But the new settings split history from training, so check both: you can now keep your chat history without letting Google use your conversations for training, and vice versa.

How does Gemini’s retention compare with ChatGPT and Claude? OpenAI says it may keep temporary chats in ChatGPT for up to 30 days, and Anthropic retains incognito chats in Claude for 30 days by default. Google’s new window of around two months is roughly double that. All three may keep some data longer for legal or safety reasons.

What should my business stop pasting into AI chatbots? Client names with contact details, addresses, health or financial information, quotes that include personal details, passwords, internal pricing, and anything covered by a confidentiality agreement. If a task needs that information, do it in your own systems or a business AI tool — not in a public chatbot.

Service areas

Serving businesses across British Columbia

On site and remote across Metro Vancouver and the Fraser Valley. Remote-first on Vancouver Island and in the Okanagan, with visits by arrangement.

Surrey & South of the Fraser

On-site and remote

Vancouver, Richmond & North Shore

On-site and remote

Burnaby, Tri-Cities & Maple Ridge

On-site and remote

Fraser Valley

On-site and remote, with planned visits further east

Vancouver Island

Remote-first, visits by arrangement

Okanagan & Thompson

Remote-first, visits by arrangement

See all service areas