AI Privacy Settings: How to Protect Your Data

Everything you type into an AI chatbot is data: your questions, the documents you upload, the photos you share and, increasingly, what the assistant remembers about you. Depending on your settings, conversations may be stored for a long time, read by human reviewers for quality and safety, or used to train future models. A few minutes in the settings menu, and some simple habits, give you far more control.

What happens to your chats

Policies differ between services and change often, but most consumer AI assistants may:

• Store your conversations in your history until you delete them, and sometimes for a period afterwards for safety and legal reasons.
• Use conversations to improve their models, unless you opt out.
• Let trained reviewers read samples of conversations to improve quality and catch abuse.
• Remember details about you across chats, if a memory feature is switched on.
• Access other data such as email, files or calendars, if you connect those apps.

Business and enterprise plans usually have stricter defaults, often excluding your data from training by contract. If you use AI at work, use the account your employer provides and follow its policy.

AI features inside other apps

AI is now built into email, office suites, phones, browsers and messaging apps, and each feature may send your content to a cloud service to process it. Check the AI settings in each app separately, not just in your main chatbot, and switch off features you don’t use. Meeting recorders and transcription tools deserve extra care, because they capture other people too; see our guide to AI meeting notes.

The five settings to check

1. Model training

Look for a setting that controls whether your conversations can be used to train or improve models. Names vary: in ChatGPT it sits under data controls, and other assistants have similar toggles in their privacy settings. For most people, switching it off costs nothing.

2. Chat history and activity

Decide how long your history is kept, and delete old conversations you no longer need. Some services let you set history to delete automatically after a period.

3. Temporary or incognito chats

Many assistants offer a temporary chat mode that isn’t saved to your history or used for training, although it may be kept briefly for safety checks. Use it for anything personal.

4. Memory and personalisation

Memory lets an assistant remember facts about you, such as your job, family or preferences. It’s convenient but builds a profile. Review what’s stored, delete anything you don’t want remembered, or switch memory off.

5. Connected apps and permissions

If you’ve connected email, cloud storage, calendars or browsers, check what access you’ve granted and remove anything you don’t use. The same goes for browser extensions and third-party apps built on AI models, which have their own privacy policies. Custom assistants and shared projects can also expose files to others; see custom AI assistants explained.

What never to paste into a chatbot

If you need help with a sensitive document, remove names and identifying details first, or use an approved work tool. Our guide to summarising PDFs with AI safely shows how.

Photos, voice and files

• Photos can reveal more than you think: faces, locations, documents in the background and, in some files, location data.
• Voice chats may be stored as audio or transcripts; check voice settings separately.
• Uploaded files may stay attached to a conversation until you delete it.

Shared links

Sharing a conversation usually creates a link that anyone can open. Shared chats have occasionally been indexed by search engines. Check what you’re sharing, and delete old shared links from your settings.

Children and students

Many AI services require users to be at least 13, and some require parental consent for under-18s. Talk to children about what not to share, and check school guidance on approved tools; see our guide to AI for students.

The most private option: run AI yourself

Open models running on your own computer keep your data on your device, with no account at all. They’re less capable than the biggest cloud models but are improving fast; see how to run AI models locally.

Frequently asked questions

Are AI chatbots safe to use?

Yes, for most everyday tasks, as long as you manage your settings and avoid sharing sensitive personal or confidential information.

Can I stop ChatGPT from training on my conversations?

Yes. Turn off model improvement in its data controls, or use a temporary chat. Similar settings exist in other major assistants.

Does deleting a chat remove it completely?

It removes it from your history. Services may keep deleted data for a limited period for safety or legal reasons, as described in their privacy policies.