AI Tool Privacy and Terms Checklist: 10 Things to Check Before You Paste Anything In
A plain-English checklist for reading an AI tool's privacy policy and terms: training on your data, retention, commercial rights, exports and more.
Most people judge an AI tool by its output and never read the policy behind it. That is understandable, but the policy decides what happens to what you type in. This checklist shows what to look for. It takes about ten minutes for a tool you plan to use seriously.
This is general guidance, not legal advice. For contracts, regulated data or anything sensitive, ask the vendor directly and involve whoever handles legal or compliance for your work.
1. Is your content used to train models?
Look for wording about "improving our services", "model training" or "product improvement". Some tools train on what you submit by default; others do not, or offer a switch. Find out which applies to your plan, because free and paid plans can differ.
2. Can you opt out, and how?
A setting you can turn off is better than a promise to ask support. Check whether the opt-out is available on your plan and whether it applies to past inputs as well as future ones.
3. How long is your data kept?
Look for a retention period and whether you can delete conversations, files and the account itself. "We may retain data as long as necessary" is not a number.
4. Who can see it?
Check whether human reviewers can read your inputs, which third-party services process them, and where data is stored. For work involving other people's information, this matters even more.
5. Who owns the output?
Terms usually say who holds rights to what the tool generates. Confirm that you can use it commercially, and whether anything restricts how you use it.
6. Are there commercial-use limits on your plan?
Some free or low-cost plans exclude commercial use. If you will use the output in client work, a product or advertising, read this line before you rely on it.
7. Can you export your work?
If you build something inside a tool, such as a project, a custom assistant or a library of content, check that you can take it with you. A tool that is easy to enter and hard to leave is a risk.
8. What are the security basics?
Look for two-factor sign-in, single sign-on for teams, role-based permissions and a security page that explains how data is protected. For businesses, ask whether the vendor can share a security summary or independent audit report.
9. What changes the terms?
Policies change. Check whether the vendor notifies you of changes and what happens to your data if you do not agree. It is worth putting a reminder in your calendar to re-check the terms of any tool you depend on.
10. What is the cancellation and refund process?
Find out how to cancel, whether it can be done online, what happens to your data after you do, and whether there is a refund window for annual plans.
Quick rules of thumb
- Never paste confidential material into a tool whose terms you have not read.
- Remove personal details from anything you do not need identified.
- Prefer a work or business plan for work data, since these often offer stronger controls than free consumer plans.
- Keep a short list of which tools your team is approved to use for which kinds of information.
How this connects to our reviews
Our Trust Score includes a privacy factor, but for most listings it is a flat default, not an individual assessment, and the page says so. Treat it as a reminder to check, not a verdict. The method is explained in how we review AI tools.
When you are ready to compare options, use our comparison tool, and run a fair trial using the 30-minute testing checklist.