People lose billions of dollars to scams every year, and reported losses keep climbing. If a message is pushing you to click, reply, or pay right now, that pressure is the tell. You do not have to figure it out alone. You can paste it into AI and have it flag the red flags in about thirty seconds, before you do anything you cannot undo.
Here is the check I use, and the safety rules that keep it from making you overconfident.
The 30-Second Scam Check
The whole method is three questions you hand to AI. It reads the message the way a suspicious friend would, and tells you what it sees.
- What are the red flags? The patterns scammers reuse. Fake urgency, a sender address that does not match the company, a link that goes somewhere odd, a strange payment request.
- What is this message trying to get me to do? The intent. Click a link, hand over a login, send money, install something.
- What should I do next? The safe move. Usually: do not click, do not reply, and verify through the real company yourself.
That is it. Red flags, intent, next step. You are not asking AI to be certain. You are asking it to slow you down and point at what looks off.
Read this before you trust it
AI is a smart second opinion. It is not a guarantee. It can sound sure and still be wrong, both ways: it can miss a real scam, and it can call a real message a scam. So a few rules matter more than the tool.
- Never click a link to check it. Not even to copy it. Copy the link as text, or skip it and go to the company’s site yourself.
- Do not paste passwords, full account numbers, or logins. You are asking AI to read words, not to sign in anywhere. Strip out anything that identifies you.
- Verify anything that matters through the official source. Call the number on the back of your card, not the number in the message. Type the website address you already know, not the one you were sent.
- Urgency is the scam. The rush is the whole game. Real companies let you take your time. If something is screaming “act now,” that alone is a reason to slow down.
Follow those four and AI makes you genuinely safer. Skip them and it can make you careless, which is worse than where you started.
I built this for my own family first
I spend my days on AI for work, and the first place I brought it home was exactly this. Someone I love got a text that looked like it came from their bank, and they almost tapped it. Now the rule in my house is simple. When anything asks for money or a password, paste it in and ask before you act.
It is not about being tech-savvy. It is about having a calm second set of eyes at the exact moment you feel the pressure to hurry. That pause is the entire point.
What it looks like filled in
Here are real red flags, what to ask AI about each, and how to tell a safe answer from a scam one.
| Red flag in the message | What to ask AI | Safe answer looks like | Scam answer looks like |
|---|---|---|---|
| ”Your account is locked. Verify now or lose access." | "Is this urgency normal for a real bank notice?” | A note that banks rarely threaten instant lockout, and a nudge to call the number on your card | You paste the message and it walks you toward the link instead of away from it |
| Sender is “service@paypa1-secure.com" | "Does this sender address match the real company?” | A flag that the domain is off (a “1” for an “l”), and a reminder to check the real site | Anything that treats the lookalike address as fine |
| A link like “bit.ly/acct-verify" | "What does this shortened link tell you, without me clicking it?” | A warning that shortened links hide the real destination, so do not click | A confident “that link is safe” with no way to actually know |
| ”Pay the fee in gift cards to release your package." | "Is a gift-card payment request a scam sign?” | A clear yes, this is a classic scam pattern, do not pay | Any answer that helps you complete the payment |
The pattern across all four: a safe answer sends you back to the official source. A scam-flavored answer, or your own wishful reading, helps you go through the message. Send yourself out, never through.
The prompt I paste
Copy this. Fill in the message with your identifying details removed. Do not include the actual link as a clickable link, just the visible text.
I got the message below and I am not sure if it is a scam.
I have removed my personal details. Do not ask me to click anything.
Tell me three things:
1. The red flags you see, and why each one is a red flag.
2. What this message is most likely trying to get me to do.
3. What I should do next, assuming I will verify with the real
company myself and will not click any link in the message.
If it looks like a scam, say so plainly. If you are not sure, say
that too, and tell me the safest next step either way.
Here is the message:
[paste the wording here, with names, account numbers, and links removed]It works on ChatGPT, Claude, or Perplexity. If you want it to check whether a company or offer is even real, use one with live web search so it can look up the actual business instead of guessing.
Common mistakes
- Clicking the link to “check” it. This is the big one. The click is often the whole attack. You cannot un-click. Copy the text or go to the site yourself.
- Trusting the AI 100 percent. It is a second opinion, not a verdict. It can be wrong in both directions. Let it flag things, then verify the real thing yourself.
- Pasting sensitive info. No passwords, no full account numbers, no logins. You are asking it to read a message, not to sign in for you.
- Acting on the urgency. The rush is the scammer’s design. The moment you feel “I have to do this right now,” that feeling is the reason to stop and check.
A quick note on the rest of your digital life
The same instinct helps beyond your inbox. When you use AI for money tasks like decoding a bill, keep the same rule: describe the situation, never paste the account number. When you hand off the small mental-load tasks of running a home, skip anything you would not want on a shared screen. And if you give your AI a saved memory so it knows your setup, be just as careful about what you let it remember.
Do one thing today
Save the prompt above somewhere you can find it fast. Notes app, pinned text, wherever. Then send it to one person in your life who would tap the link without thinking. That single share does more than reading this ever will.