Don't Put This into AI — Safety Rules
The moment you type information into AI, have you ever stopped to think about where it goes?
It's so convenient that we paste things in without thinking — but there are things you should never enter. What to watch out for, and how to use AI safely — here's the clean rundown. 👇
The bottom line (30-second version)
| This kind of info | Don't enter it |
|---|---|
| 🆔 ID / passport / account / card numbers | ❌ Never |
| 🏢 Company secrets, unreleased files, customer databases | ❌ Never |
| 👥 Other people's personal data (contacts, addresses) | ❌ Never |
| 🔑 Passwords, verification codes, API keys | ❌ Never |
Why, and how to stay safe? 👇
🚫 What you must never enter
1. Anything that identifies you — ID number, passport number, account and card numbers, your exact home address.
2. Company secrets — unreleased results, contract text, source code, customer lists. Feeding company material into an outside AI can amount to a data leak.
3. Other people's personal data — a colleague's or client's name, contact, or address. If it isn't yours, be even more careful.
4. Passwords and credentials — passwords, SMS verification codes, API keys. These are things you wouldn't hand to another person either.
❓ Why it's risky
- What you enter may be used to train the AI or stored on a server (it varies by service and settings).
- Once information is out, it's hard to take back.
- Free apps and AI apps from unknown sources are especially opaque about how they handle your data.
💡 One test is all you need — "Would I be fine showing this to a stranger?" If not, don't enter it.
Now, how to use AI 'safely anyway.' 👇
🛡️ How to use AI safely
1. Anonymize before you type — turn names into 'Person A', companies into 'Company B', and round the numbers off. Give it the context and AI can still help plenty.
2. Turn off training and history — most AIs offer a setting to "not use my conversations for training" or to stop saving history. Switch it on before any sensitive work.
3. Check your company's policy — for work, first confirm whether it's an AI your company allows and whether any internal rules apply.
4. Official apps only — steer clear of 'free AI' apps and extensions from unknown sources.
Leaving you with an 'anonymize' prompt you can use right away. 👇
By situation — this is OK, this isn't
The confusing real-life cases, sorted out at a glance.
| Situation | Verdict |
|---|---|
| Polishing my own résumé | ⚠️ Leave out name and contact — content only |
| Summarizing the text of a company contract | ❌ Off-limits unless it's an approved tool |
| Asking about medical test results | ⚠️ Remove names and numbers, and only for reference |
| Summarizing an unfamiliar word or article | ✅ Go ahead, no problem |
| Uploading a family photo to edit | ⚠️ Watch faces and location data; don't spread it publicly |
💡 One rule covers it — "Would it be fine to show this to someone who doesn't know it?" If you're unsure, swap in anonymized stand-ins (Person A, Company B, 123).
Copy-paste — do this instead of sensitive info
Help me with the situation below. (Real names, company names, and numbers are swapped for A, B, 123, etc.)
"An email to contact B at Company A, politely explaining the delayed project schedule."
→ I'll fill the real names and figures into the finished answer myself.
FAQ
Q. Is paid AI safe? It's better, but never 100%. Paid or free, don't feed it sensitive information.
Q. I already entered something. Delete the chat history in that service's settings, and immediately change any exposed passwords.
Q. So I can't summarize company documents? You can, if it's a company-approved tool. If not, anonymize just the essentials and summarize that.
Read next
- 👉 When AI Keeps Getting It Wrong — How to Reduce Hallucinations
- 👉 10 Hidden AI Chatbot Features You're Missing
Hashtags: #AIsecurity #personaldata #privacy #AIcautions #safetytips
