Privacy settings and sensitive data

When you finish, you can adjust the settings in your tools and sort information into three safety levels.

Find the controls

Most AI tools have a settings area with options about chat history, saving, and whether your conversations may be used to improve the product. The names and places change from product to product and from month to month, so look for words like privacy, data controls, history and training.

Make this a ten-minute job today. Open each AI tool you use, find those words, and read what each switch does. Pick the setting you are comfortable with. Some tools also offer a temporary or private chat mode for one-off questions. Check the product's current help pages to see what yours offers.

Three levels of information

Sort what you might type into three groups. Green is safe: general questions, public facts, your own made-up examples. Yellow needs cleaning first: anything with names, places, dates or numbers that could point to a real person. Red is never pasted into a consumer tool: passwords, bank and card numbers, government ID numbers, medical or legal files about real people, and secrets that belong to your employer.

The sorting takes seconds once it is a habit. Ask yourself one question: if this text appeared on a public noticeboard, would someone be harmed or embarrassed? If yes, it is yellow or red.

Cleaning yellow information

You can keep the useful part of a task and remove the identity. Replace names with roles, such as [CUSTOMER] or [NEIGHBOR]. Replace exact figures with ranges or placeholders. Remove addresses, account numbers and anything unusual that only one person would match.

A weak prompt: "Summarize this complaint from Mr. Okafor at 14 Lagos Road about order 88231." A strong prompt: "Summarize this complaint from a customer about a late order. I removed the details." The summary will be just as useful.

Rules for regulated work

If you work in health, law, finance, education, counseling, insurance or government, extra rules apply. Never put patient, client, student or citizen identifying information into a consumer AI tool unless your employer has approved that tool for that data. Follow your employer's policy and the rules of your licensing or regulatory body.

Health-privacy laws such as HIPAA in the US, and similar laws in many other countries, exist to protect people in these situations. This course is not legal advice. When in doubt, do not paste it, and ask your privacy officer or manager.

Practice now

List the AI tools you use. For each one, write down the setting that controls chat history or training, and whether you want it on or off.
Write ten things you often type into AI tools and label each green, yellow or red. Rewrite two yellow items so they become green.

Use any AI assistant you have. Do not paste private, client or patient information into it.

Check yourself

1. Which item is red, never pasted into a consumer AI tool?

2. What is a good way to clean a yellow request?

3. A nurse wants help with a patient's note. What comes first?

Ask the AI tutor

Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.

Next lesson: Bias and fairness

AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.