5-minute lesson

What should I never put into AI?

Quinn asks a practical question for work, home, and everyday AI use: “What should I remove before I paste this?”

Quinn thinking through a privacy question

Quinn asks

The question

“Can I use AI for this without sharing something I should keep private?”

Professor Tutus explains

Use the least sensitive version that still works.

AI tools are useful because they can work with context. But context can include names, addresses, customer details, medical information, finances, private conversations, company files, or details that belong to someone else.

The habit is simple: pause before you paste. Ask whether the AI really needs the real details, or whether it can help with a cleaned-up version of the problem.

You are not trying to be scared of AI. You are trying to be careful with information that would be hard to take back once it has been shared.

ASTUTE lens

Three checks before you paste.

Aware

What private, personal, or work-specific details are inside this prompt?

Trustworthy

Do I know where this information may go, be stored, or be reviewed?

Thoughtful

Would someone else be surprised or uncomfortable if I pasted this?

Safer prompt habits

Keep the shape. Remove the sensitive parts.

Instead of pasting a full document

Remove names, numbers, addresses, client details, and anything confidential. Ask about the structure or pattern instead.

Instead of sharing a private situation

Describe the decision you are trying to make without exposing someone else’s personal details.

Instead of uploading work data

Use a short fake example that keeps the shape of the problem but strips out real company, customer, or financial information.

Quinn studies

Use the “privacy pause.”

Before you paste, ask: “Could I get the same help with names, numbers, private details, and sensitive context removed?” If yes, use the safer version.

Quick check

Can Quinn pass the privacy pause?

1 Which prompt is safest to put into an AI tool?

A fake example that keeps the structure of the problem but removes real names and private details. Correct. A fake or cleaned example lets you get help without exposing information that does not need to be shared. A full customer email with name, phone number, order number, and complaint details. Not quite. That includes personal and customer-specific information that should be removed first. A private medical, legal, or financial document because AI can summarize it quickly. Not quite. Sensitive documents need extra care, and sometimes should not be pasted into a general AI tool at all.

2 What should you do before pasting work information into AI?

Check the tool rules, remove sensitive details, and ask whether the same help is possible with a safer example. Correct. The safest habit is to pause, clean the prompt, and use the least sensitive version that still gets the job done. Paste everything so the AI has maximum context. Not quite. More context is not always better if that context exposes private, confidential, or unnecessary information. Assume it is fine if the tool is popular. Not quite. Popular tools can still have rules, settings, and risks you need to understand.

3 Which question is the best “privacy pause”?

Could I get the same help if I removed or changed the sensitive details? Correct. That question keeps the usefulness while reducing unnecessary exposure. Will the AI answer faster if I paste the whole thing? Not quite. Speed matters less than whether the information should be shared in the first place. Does the prompt feel interesting? Not quite. Interesting is not the privacy test. Necessity and sensitivity are the test.