Here’s the thing about chatbots like ChatGPT: they feel weirdly private. It’s just you and the screen. No one’s watching. So it’s tempting to just… paste in whatever you need help with. Your doctor’s note. Your bank statement. Your login info.
Don’t.
Anything you type into an AI chatbot isn’t automatically deleted or locked away somewhere safe. Researchers writing for the Stanford Human-Centered AI Institute note that the chatbot can draw inferences from what you share — and the algorithm may decide you fit a classification as a “health-vulnerable individual.” That determination can filter into the developer’s broader ecosystem. You start seeing targeted ads for medications. It gets uncomfortable fast.
The Mozilla Foundation points out that even when a chatbot says your prompts won’t be used to train the AI, information about you — personal or not — may still be used in other ways. The fine print matters, and most of us don’t read it.
So what specifically should you never type in? The list is pretty clear:
— Your Social Security number, passport details, or driver’s license info. If there’s ever a data breach, hackers can use that to steal your identity or commit financial fraud. The damage can follow you for years.
— Usernames and passwords. Any of them. Ever.
— Credit card numbers, bank accounts, or tax documents.
— Medical stuff — test results, prescriptions, your health history. Once you type it, it’s no longer just yours.
The rule of thumb is simple: if you wouldn’t hand that information to a stranger on the street, don’t hand it to a chatbot either. Ask for help with your essay. Brainstorm ideas. Get a recipe. Just keep the sensitive stuff off the keyboard.