Step-by-step

🛡️ AI skills

For everyday people

📘 Step-by-step beginner 🏠 Everyday life

How to check if a suspicious message is a scam by pasting it into an AI

Q: What mistakes should I avoid?

Pasting real passwords or full card numbers. The AI doesn't need them, and storing them in a chat creates risk if your account is ever hacked. Fix: redact first (Step 2), or describe the message in your own words instead of pasting it. Trusting the AI's verdict without checking. AI assistants sometimes miss obvious scams or flag clean messages by mistake. Fix: always cross-check the specif

A 5-minute way to get a fast second opinion on a sketchy email or text before you click anything

6 min read·September 17, 2026

This guide was written by AI. It passed automated fact and quality checks, and our editor spot-checks published guides every day.

See AI World HQ first on GoogleAdd us as a preferred source and our stories show up higher in Google’s Top Stories.Add on Google

Scam messages have gotten good. They use real company logos, your real name, sometimes even your real address. Before you click anything, copy the message into an AI assistant and ask for a second pair of eyes. This guide walks you through that in six steps — even if you've never used an AI assistant before. One honest warning up front: the AI is a useful first check, not a verdict, and it cannot actually click the link or call the company for you. You still have to do that part.

💡 Tip: tap a step’s number when you finish it — a green tick appears and your browser remembers how far you got.

✅ Before you start

An AI assistant you can sign into — ChatGPT, Gemini, Claude, Microsoft Copilot, and Le Chat all work, and their free plans are fine for this.
The suspicious message itself, ready to copy (an email, a text, a chat screenshot, or a DM).
About 5 minutes in total.
Important: never paste real passwords, full card numbers, or government ID numbers, even if the message contains them — Step 2 shows you how to handle that.

Step 1 of 6

Open a fresh chat in any AI assistant

Open the website or app of any AI assistant — ChatGPT, Gemini, Claude, Microsoft Copilot, Le Chat, or whichever you already use. Start a brand-new conversation rather than dropping the message into an older thread. On a phone, the app icon usually looks like a small chat bubble or a starry logo. On a computer, just go to the assistant's homepage and click into the empty chat box. A clean thread means the AI won't get distracted by older context and will focus entirely on your message. If you don't already have an account, the free sign-up usually takes under a minute and only needs an email address.

💬 Exampleopen ChatGPT, Gemini, Claude, Microsoft Copilot, or Le Chat — whichever one you already have an account on.

✅ You'll know it worked when

you see an empty chat box with a placeholder like "Message…" or "Ask anything" and a send button (often an arrow or paper-plane icon).

Next step ↓

Step 2 of 6

Redact personal details before you paste

Before you copy the message, take 30 seconds to hide any sensitive bits it contains. Cover real passwords, full credit card numbers, your Social Security or national ID number, your home address, and any one-time login codes. A quick search-and-replace in any notes app works fine. Think of it like covering your hand on a debit-card PIN pad at the store — the AI doesn't need those details to spot a scam, and fewer real numbers in the chat mean less risk if your account is ever hacked. If the message has no personal details at all, you can skip ahead to Step 3.

💬 Exampleturn "my SSN is 123-45-6789, please call 555-0123" into "my SSN is [REDACTED], please call [REDACTED]".

✅ You'll know it worked when

the message still reads naturally but no real numbers, codes, or addresses remain.

Next step ↓

Redact personal details before you paste

Step 3 of 6

Paste the message and ask the AI to flag red flags

Tap or click inside the empty chat box and paste the redacted message. Then add a short prompt asking the assistant to look at it as if it were a security expert. Press the send button (usually an arrow ↗ or paper-plane icon, but apps vary). After a few seconds the AI will start typing a reply — usually a short list of warning signs plus a one-line verdict like "this looks like a phishing attempt" (a fake message designed to steal your login or money) or "this appears legitimate, but verify the sender". If nothing appears after a few seconds, look for a small "stop" or square icon — the response sometimes needs a moment to begin. Apps change often, so if your screen looks a little different, just look for any button that submits your message.

💬 Try typing thisType your own details into the dashed boxes, then copy."Here's a message I just received. Can you act as a security expert and tell me whether it looks like a scam? List the specific red flags you see, and tell me whether I should click any links in it. Message: "➤

✅ You'll know it worked when

you see a written response that names specific warning signs, such as urgency, suspicious links, an odd sender, or requests for personal info.

Next step ↓

✏️ Sketch — your screen may look a little different

Step 4 of 6

Read the AI's answer with a critical eye

Treat the AI's reply like a friend's opinion, not a doctor's diagnosis. Read each red flag it names and ask yourself: does this actually match the message I received? Many scams lean on urgency ("act in the next 24 hours"), impersonate a brand you trust, include a link that almost looks right, ask you to click or log in, or demand payment in gift cards or cryptocurrency. If the AI flags something that isn't actually in your message, ignore that point — assistants sometimes guess wrong, and you are still the expert on your own inbox. Conversely, if something feels off to you but the AI didn't mention it, trust your gut and move to the next step. Be especially wary if the AI says "this looks fine" but the message asks you to click a link, send money, or share a code — that's the moment to get a second human opinion.

💬 Exampleif the AI says "the sender domain doesn't match the company", check that yourself — does the "from" address really end in the company's real website, or in something like "support-service.example.com"?

✅ You'll know it worked when

you can repeat back, in your own words, two or three specific reasons the AI gave for its verdict.

Next step ↓

✏️ Sketch — your screen may look a little different

Step 5 of 6

Verify any links or sender names yourself

Don't click anything yet. If the message contains a link, hover over it on a computer or long-press it on a phone — you'll see the real destination. Real companies use their own short domain (the brand's actual homepage), while scam links often hide behind words like "secure-login" or "verify-account" on a totally different address. The simple rule: if the link doesn't end on the brand's real site, it's almost certainly fake. If the message claims to be from a person you know, contact that person through a channel you already trust — a phone call or a separate chat — and never reply to the suspicious message itself. If it claims to be from a company like your bank, look up the official number on their real website (type the address yourself, don't use any number inside the message) and ask them directly.

💬 Examplea "USPS" tracking link points to `usps-tracking.example.com/pay-now` — that is NOT usps.com, so it's a fake.

✅ You'll know it worked when

you have either confirmed the message is genuine through a trusted channel, or you've concluded it's a scam.

Next step ↓

✏️ Sketch — your screen may look a little different

Step 6 of 6

Decide what to do, and don't act in a rush

Now choose one of three actions. If it's a scam: delete it, block the sender, and — if it pretended to be a company — report it to that company through their real website. If it's genuine but unexpected: still don't click links inside it; go to the company's real site yourself and log in there. If you're not sure: leave it alone for an hour, then re-read it cold; urgency is the scammer's favorite tool, and a slow second look usually reveals more. Tell a friend or family member if you're worried — a second human pair of eyes still catches things AI misses, especially with voice or video scams. The whole point of using an AI here is to slow down, not to rush.

💬 Exampletext a family member — "Hey, I just got a weird message about my Amazon account, does this sound legit to you? [paste]"

✅ You'll know it worked when

you take one concrete action in the next few minutes — delete, report, or verify through a trusted channel — instead of leaving the message sitting unread in your inbox.

Decide what to do, and don't act in a rush

⚠️ Common mistakes

Pasting real passwords or full card numbers. The AI doesn't need them, and storing them in a chat creates risk if your account is ever hacked. Fix: redact first (Step 2), or describe the message in your own words instead of pasting it.
Trusting the AI's verdict without checking. AI assistants sometimes miss obvious scams or flag clean messages by mistake. Fix: always cross-check the specific red flags it names against the actual message (Step 4) and verify with the real company (Step 5).
Clicking the suspicious link "just to see what happens". That's exactly what the scammer wants. Fix: hover or long-press to read the real URL, and never log in through a link inside an unexpected message — go to the company's website yourself.

🚀 Try it now

Find one message in your inbox you've been unsure about, redact any personal details, paste it into your AI assistant, and ask it to flag the red flags. Five minutes, one fresh chat, one safer decision.

🎯 Next guideHow to Fact-Check a Suspicious Text or Link Before Forwarding ItUse any AI assistant to spot warning signs, build a reusable checklist, and decide what to share with family, in about ten minutes.

❓ Quick questions

How long does this take?

About 6 minutes — the guide has 6 steps, and you can tick each one off as you go.

Do I need to prepare anything?

An AI assistant you can sign into — ChatGPT, Gemini, Claude, Microsoft Copilot, and Le Chat all work, and their free plans are fine for this.
The suspicious message itself, ready to copy (an email, a text, a chat screenshot, or a DM).
About 5 minutes in total.
Important: never paste real passwords, full card numbers, or government ID numbers, even if the message contains them — Step 2 shows you how to handle that.

What mistakes should I avoid?

Pasting real passwords or full card numbers. The AI doesn't need them, and storing them in a chat creates risk if your account is ever hacked. Fix: redact first (Step 2), or describe the message in your own words instead of pasting it.
Trusting the AI's verdict without checking. AI assistants sometimes miss obvious scams or flag clean messages by mistake. Fix: always cross-check the specific red flags it names against the actual message (Step 4) and verify with the real company (Step 5).
Clicking the suspicious link "just to see what happens". That's exactly what the scammer wants. Fix: hover or long-press to read the real URL, and never log in through a link inside an unexpected message — go to the company's website yourself.

Keep reading

📘 How to Spot a Phishing Email or Scam Text Before You Click Paste any suspicious message into an AI assistant with a fraud-detector prompt and it will flag the red flags your eyes missed 📘 How to Spot an AI Scam Call, Fake Email, or Deepfake Voice Today Three AI-powered scams happening right now and the one-second checks that catch each one 📘 How to Fact-Check a Suspicious Text or Link Before Forwarding It Use any AI assistant to spot warning signs, build a reusable checklist, and decide what to share with family, in about ten minutes. 📘 Translate a Foreign-Language Message in Messenger with Meta AI Get a plain-English translation of any message in seconds, without copy-pasting or switching to another app 📰 Google DeepMind's New On-Device AI Is Built to Catch Chat Scams A partnership with Singapore's GovTech uses lightweight Gemma models to spot scam patterns in private messages — without sending your data anywhere. 📘 How to quickly fact‑check any AI answer in under one minute Learn a simple, step‑by‑step method to verify AI‑generated information fast, using source requests and quick cross‑checks so you can spot mistakes before you act.

📬 The week’s AI, in your inbox

One friendly email every Sunday — the 5 stories that mattered, in plain English. No spam, unsubscribe anytime.

Was this helpful?

✦ Generated by AI in AI World HQ's automated newsroom, written in plain language. Checked by automated fact and quality gates, and our editor spot-checks published guides every day. Spot a mistake? Use the buttons above.

📘 Prefer it in one file? Scams, Deepfakes and Privacy: The AI Safety Pack (8 guides, free printable PDF)

← Back to all stories