How to stop AI from using your information

There's a very common confusion about this, and it's worth clearing up before you touch any setting: the line isn't between the free plan and the paid one. It's between a personal account and a company account.

A paid personal plan, the twenty-euros-a-month kind, is still a personal account. By default, what you type there can be used to train the model, and to prevent that you have to go into settings and switch it off yourself. Company accounts come the other way around out of the box: team plans, enterprise plans and API access don't train on your data unless somebody deliberately turns it on.

So if your team uses AI every day on paid personal accounts, convinced that paying covers you, the switch is in the wrong position. Here's how to move it.

ChatGPT

Settings, then Data controls, and switch off "Improve the model for everyone".

It's a separate toggle from your chat history. You can keep saving your conversations and still have training switched off, which is what most people actually want. There's also temporary chat, which isn't used for training and isn't saved to your history, for the one-off conversations you'd rather leave no trace.

Team and Enterprise plans and API access are already excluded from training by default, with nothing for you to change.

Claude

The direct route is claude.ai/settings/data-privacy-controls. Through the menu: your name at the bottom left, Settings, Privacy, and switch off "Help improve Claude".

Here the setting changes two things at once. It stops your new conversations being used for training, and it drops retention from five years to thirty days. That's the biggest gap of the three tools, and it's the reason this is the switch most worth moving.

What doesn't change: content flagged by the safety systems still gets reviewed even with it off. There's no opting out of that, and it's reasonable that there isn't.

Claude for Work, team plans and the API run on different commercial terms and aren't part of training.

Gemini

This one works differently from the other two, and it's where most people get it wrong.

In ChatGPT and Claude, training has its own toggle, separate from your history. In Gemini they're tied together. The control is called "Gemini Apps Activity" and lives in myaccount.google.com, under Data and privacy. Turning it off cuts your history, human review and the use of your conversations to train Google's models all at once.

Two things worth knowing. Google still keeps conversations for up to 72 hours even with it off, because it needs them to run the service. And if a human reviewer did see a conversation of yours earlier, that one is kept for up to three years, disconnected from your account.

Google Workspace business, enterprise and education accounts are contractually excluded from training the general models.

What this doesn't fix

Three things, worth being clear about so you don't relax more than you should.

It only works going forward. Whatever you wrote before switching it off may already be inside the process, and that train doesn't come back.

It doesn't turn the tool into a safe. Not training on your data doesn't mean your data exists nowhere: it still travels across the internet and lives on a server that isn't yours.

And it doesn't excuse you from thinking. Does it mean you can't give it a client's name or a project's details? No. It means the genuinely confidential stuff, the kind you wouldn't want to see in an email forwarded by mistake, is probably better handled by a person rather than an outside service.