wasl.tech

Google Splits Gemini Privacy Settings: Keep Your History Without Feeding AI Training

Picture this: it is past midnight, and you are asking Gemini to help you draft a difficult message to your boss. Or to make sense of a medical report. Or to untangle a family budget that you would rather nobody else saw.

You finish, you close the app, and a small worry follows you to bed. Where did those words go? Are they sitting in a history list you can search later — or are they quietly becoming training data for the next version of Google’s AI?

Until now, Gemini forced you to pick one answer for both questions. That is about to change. Google has announced it is splitting its Gemini privacy settings in two, so you can finally keep your chat history without handing your conversations over for AI training.

There is, however, a catch hiding in the small print — and it is a big one.

What the Gemini Privacy Split Actually Changes

Google announced the change on November 16, 2026, updating its Gemini Apps Privacy Hub the same day. The old setting was called Keep Activity, and it did two jobs at once. It decided whether your chats were saved in your history, and it decided whether Google could use them to train its generative AI models.

That bundling created an awkward trade-off. If you wanted a searchable record of your conversations, you had to accept that those conversations could also become training material. If you wanted to keep your words out of training, you had to give up your history entirely.

The new setup separates the two choices. One setting, called History, decides whether your chats are saved and for how long. A second, separate setting decides whether the content of your chats can be used to train Google’s generative AI models.

“New settings let you keep your Gemini chat history, while opting out of generative AI model training,” Google said in its announcement, according to 9to5Google, which first reported the change.

The company says it will start rolling out the new controls globally in the Gemini app over the coming weeks, and it is emailing users about the switch.

How the Two New Gemini Privacy Settings Work

The History setting is the straightforward one. When it is on, your chats are saved in your history, and Google deletes them automatically after 18 months by default. You can change that period to three months or 36 months, or switch auto-delete off altogether.

When History is off, your future chats will not appear in your history at all.

The training setting is the one that matters most for privacy. When it is on, Google uses the content of your future chats to train its generative AI models.

That covers more than the words you type: prompts and audio transcripts, photos, files and other attachments, screens you ask about, Gemini’s responses, tasks and tool calls, and data from apps you connect to Gemini.

That is a wide net — and it is why the new opt-out matters.

When the training setting is off, your future chats will not be used to train generative AI models — unless you provide feedback on a response. There is one dependency worth knowing: the training setting can only be switched on if History is on. Training needs history; history no longer needs training.

If you already have Keep Activity switched off today, Google says both new settings will start off when the change reaches your account.

And if you pay for Google AI Pro or Ultra, do not assume that buys you out — a paid plan still does not opt you out of training on its own. You have to flip the switch yourself.

The Two-Month Catch in Gemini’s New Privacy Rules

Here is the part Google is not shouting about. Turning History off used to mean your chats were gone quickly. Under the old rules, chats with Keep Activity off were retained with your account for 72 hours. The new version of the privacy hub says something very different.

“Even if Gemini History is turned off, future chats will still be saved for around two months and used to respond to you, process your feedback, and protect Google, its users, and the public, with help from human reviewers,” the hub now states.

That is a jump from three days to roughly sixty.

Google says the longer window exists partly so chats are available in case of a potential system failure, and to respond to you and process feedback. Temporary chats get the same treatment — they are kept for around two months too, though Google says they are not used to train its models.

There is another retention rule that survives the split. If you send feedback on a Gemini response while History is off, Google also collects the last 24 hours of your chats, including attachments.

And chats picked for human review are separated from your account and can be kept for up to three years, even if you delete them later or turn a setting off.

Why Google Changed Its Gemini Privacy Settings Now

The timing is hard to miss. Just last month, Google quietly added a note to a 2023 blog post saying Gemini data “can be used to train generative AI models” and “can also be subject to human review in limited circumstances.”

That discovery, which spread widely after reaching the top of Hacker News, put a spotlight on how much of your Gemini activity Google could see.

It was a reminder that the old switch hid a lot behind one toggle.

Google’s privacy hub now says a subset of data from connected apps — including Gmail and Drive — can be reviewed by human reviewers, including contractors.

Excerpts, summaries or inferences are reviewed, and sometimes the full text. Reviewed data can be kept for up to three years, and deleting it does not remove it from that review store.

We covered that earlier revelation in our report on Gemini’s human reviewers reading Gmail and Drive files.

That earlier story is what made the old all-or-nothing switch look so dated. Splitting the controls answers the most obvious complaint — while the two-month retention quietly answers a different one, about how long Google keeps your words even when you say no.

The change also lands in a week when Google has been moving fast on Gemini. The company made Gemini Tasks free in six countries just yesterday, and its Gemini 4 Carbon model arrived only last week.

Gemini is becoming the front door to more of Google’s services, which makes the question of what happens to your conversations more important, not less.

What to Check When the Gemini Privacy Update Arrives

The new settings are rolling out over the coming weeks, so they may not be on your phone yet — Notebookcheck found a German account still showing the old setting on November 17. When the update lands, here is what to look for.

First, open the Gemini app’s settings and look for the two new controls. If you want your history saved but your chats kept out of training, turn History on and leave the training setting off. That combination did not exist before, and it is the whole point of the change.

Second, remember that the training setting is off by default, according to Google’s history page. If you never touch it, your future chats should not be used for training. But check anyway — defaults have a habit of changing, and this is the setting that decides whether your words feed Google’s models.

Third, do not mistake History off for deletion. Your chats will still be held for around two months, and feedback or human review can extend that to three years. If you share something sensitive with Gemini, assume it exists on Google’s servers for weeks at minimum, whatever your settings say.

What Happens Next for Gemini Privacy

The split is a genuine improvement, and it is fair to say so. For the first time, Gemini users can have both a useful, searchable history and a real opt-out from training. That is the arrangement many people assumed they already had.

But the two-month retention is the part that will keep privacy advocates talking. Google has traded a short, clear deletion window for a longer, fuzzier one, in the same breath as giving users more control.

Whether that trade is acceptable depends on what you use Gemini for — and how much you trust that a chat you chose not to save is truly out of reach.

The rollout will take weeks, and the real test will come when millions of people open their settings and see the two switches for the first time.

How many turn training off? How many notice the two-month footnote? Until then, the safest move is the simplest one: check your settings the moment the update arrives, and decide for yourself what your words are worth.

Source: Notebookcheck, reporting on Google’s Gemini Apps Privacy Hub update of November 16, 2026.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top