How to Spot Scam Messages with ChatGPT | OpenAI Academy

By OpenAI

Categories: AI, Product

Summary

Scam messages exploit fear, urgency, and relationships to bypass critical thinking—but ChatGPT can serve as a reliable second opinion. The key defense: pause when pressured, ask AI to spot red flags, then verify through official channels independently.

Key Takeaways

  1. Use ChatGPT as a safety filter before engaging with suspicious messages—copy text without clicking links or calling numbers, then ask it to identify pressure tactics, strange wording, and suspicious requests.
  2. Scam messages use three primary manipulation tactics: fear, urgency, and relationship leverage to force fast decisions before rational evaluation occurs.
  3. Always verify independently through official channels: open banking apps yourself, call card numbers on back of card, type company websites directly—never use links or numbers from the suspicious message.
  4. Create a simple verification checklist: Pause, Look Carefully, Verify Somewhere Else, Ask Someone Trusted—reduces cognitive overload when facing pressure tactics.
  5. The psychological trigger for scams is urgency—when a message pushes you to act in a hurry, that's the signal to deliberately slow down and get a second opinion.

Related topics

Transcript Excerpt

Have you ever received an email or text like this? Your package cannot be delivered. Pay a small fee now. Suspicious activity was found in your bank account. Click here immediately. Hi Grandma, this is my new number. I'm in trouble and need money. Scam messages often use fear, urgency, or familiar relationship to make you act before you have time to think. When a message pushes you to act in a hurry, that is a good reason to slow down. ChatGPT can be a second set of eyes for you, helping you spot warning signs that you might be dealing with a scam. I'm going to copy the text into ChatGPT. I am not opening the link. I am not calling the number in the message. I am just asking for a second look. Point out warning signs in this message. Do not open links. Tell me how to verify safely through …

More from OpenAI