How to keep your private information safe when using AI

You can use AI tools safely by sharing less, checking your settings, and treating every chat like a conversation in a public place. Most of the risk comes from what you type, so the habits below matter more than any technical trick. Here are five simple steps.

Step 1: Leave out the details that identify you

An AI chatbot can help with almost any question without knowing your full name, address, or account numbers. Before you press enter, ask yourself: would this still make sense with those details removed?

Instead of typing "I am Mary Jones at 14 Elm Street and I need a letter to my landlord," try this:

"Write a polite letter from a tenant to a landlord asking for a broken heater to be fixed. Leave blanks for names and addresses."

You then fill in the blanks yourself, outside the chat.

Step 2: Use placeholders

When your question involves real people, swap in stand-ins such as "my neighbor" or "Person A." Try this:

"My friend, whom I will call Sam, keeps cancelling plans. Help me write a kind message asking if everything is okay."

Step 3: Look at your privacy settings

Many AI tools let you choose whether your chats may be used to improve the product, and most let you delete past chats. Settings change often, so open the menu, find "Privacy" or "Data controls," and read what is on offer. If you are unsure, pick the more private option.

A helpful prompt for learning your way around:

"Explain in plain English what a privacy setting called 'data controls' usually does in an AI chat app. I will check the real settings myself."

Step 4: Use a strong, unique password and two-step sign-in

Your AI account holds your chat history, so protect it like an email account. Use a password you do not use anywhere else, and turn on two-step verification if it is offered. A password manager can help. Never type a password into the chat itself.

Step 5: Check before you trust

AI can sound confident and still be wrong. Do not rely on it alone for anything that affects your health, legal matters, or money. Use it to understand the basics and to prepare questions, then confirm with a doctor, a lawyer, your bank, or an official source such as Medicare.gov.

Try this:

"I have a question about a medication I take. Give me general background and a list of questions to ask my pharmacist. Remind me you are not a doctor."

What to watch out for

• Look-alike websites and apps. Open AI tools from the official company website or app store listing, not from an ad or a link in a message.
• Pop-ups asking you to "verify" your account. Real tools will not ask for your bank details in a chat.
• Uploading documents. A photo of a bill or form may contain numbers and addresses. Cover them up first, or type only the part you need help with.
• Shared computers. Sign out when you finish, especially on a library or family device.
• Feeling rushed. Anyone, or any message, pushing you to act fast and share details is a warning sign.

A good rule of thumb

If you would not read it aloud to a room of strangers, do not type it into a chatbot. When in doubt, leave it out, and ask your question in a general way.

More guides

• Is AI Safe to Use? What to Share and What to Keep Private
• Fun Activities With Grandchildren Using AI
• AI for Healthcare Workers: Safe Uses for Admin Tasks