To detect an AI scam, stop judging how convincing a message looks and start verifying who sent it, through a channel you already trust. AI has made fake emails, voices and videos polished enough that spotting flaws is no longer reliable. The dependable checks are simple: pause, verify identity independently, inspect where links and payments lead, and never share codes or passwords on request.
For an overview of the scam types themselves, such as voice cloning, deepfake executives and fake job offers, see our guide to AI scams. This article is the practical checklist.
Why "spot the typo" no longer works
For years, people were taught to look for bad grammar, strange logos and generic greetings. Those clues still catch lazy scams, but generative AI writes fluent, personalized text in any language, and image tools produce clean logos and documents.
Voice and video deepfakes add another layer. You may hear what sounds like a family member or see a familiar face on a video call. Your senses, under stress and over a weak connection, are not a reliable detector.
So the strategy shifts from inspecting the content to verifying the source.
The four-step verification routine
Use this routine for any unexpected message, call or video that asks you to do something.
Step 1: Pause
Scams depend on speed. A short pause is the single most effective defense. Tell yourself, or the caller, that you will check and get back to them. A legitimate contact will understand. A scammer will push harder, and that pressure is itself a signal.
Step 2: Identify the request
Ask: what exactly does this message want me to do? Be suspicious of any request to:
- Pay money, especially by gift card, cryptocurrency, wire transfer or payment app
- Share a password, one-time code or PIN
- Install software or give someone remote access to your device
- Send ID documents, bank details or tax information
- Keep the request secret or skip normal procedures
Step 3: Verify through a separate channel
Contact the person or organization through a channel you find yourself, not one provided in the message.
- Bank or service: open the official app or type the web address you know. Call the number on the back of your card.
- Employer or manager: message them on your usual work channel or call a number from the company directory.
- Family member: hang up and call their saved number. Ask another relative to reach them.
- University or school: log into the official portal or contact the department through its listed details.
Step 4: Inspect links, senders and payments
If you still need to look closer:
- Hover over or long-press links to see the real address. Check the domain carefully for extra words, swapped letters or unusual endings.
- Look at the full sender address, not just the display name.
- Check whether a payment recipient's name matches the organization they claim to be.
- Search the company name plus "scam" or the exact wording of the message.
Checking specific formats
Different scam formats need slightly different checks.
Emails and text messages
- Do not click links in unexpected messages. Go to the service directly instead.
- Treat attachments from unknown senders as risky, especially files that ask you to enable content.
- Be suspicious of messages that mention a real recent event (a delivery, a login, a tuition payment), since scammers time messages to look relevant.
Phone calls and voice notes
- Do not trust caller ID; numbers can be spoofed.
- Ask a question only the real person would know, and not something visible on social media.
- Hang up and call back on a known number. If it was real, nothing is lost.
Video calls and video clips
- Ask the person to do something unplanned: turn their head fully, wave a hand across their face, or hold up a specific object. Some deepfakes struggle with this, though better tools are improving.
- For any financial request on a video call, confirm through a separate channel and follow normal approval rules, regardless of who appears to be asking.
- For celebrity endorsement clips, check the person's official accounts or reputable news. Real endorsements rarely appear only in an ad.
Websites, shops and listings
- Look up how long the domain has existed and whether the business has an independent track record.
- Read reviews on independent platforms rather than on the site itself.
- Use payment methods with buyer protection. Avoid sites that only accept bank transfer, crypto or gift cards.
- For rentals, never pay a deposit before seeing the property or verifying the landlord.
Job, internship and scholarship offers
- Check the company's official careers page for the role.
- Expect a real interview with a person. Chat-only "interviews" are a warning sign.
- Never pay for training, equipment or a background check to get a job.
What about AI detection tools?
It is natural to wonder whether an AI detector can catch a scam. Usually it cannot, and it is worth being clear about why.
- A text detector estimates whether writing resembles AI output. Many legitimate companies use AI to draft emails, and many scams are written by people. "AI-written" and "scam" are not the same thing.
- Deepfake detection tools for images and video exist, but results vary, and they are not a substitute for verifying identity.
- A fact-checking tool can help you check a claim inside a message, such as whether a law, policy or news event is real, but it cannot confirm who sent the message.
The reliable signal is always the same: does this request make sense, and can you confirm it through a channel you trust?
A worked example
A student receives a text: "Your tuition payment failed. Pay the outstanding balance within 2 hours to keep your course registration," followed by a link. The message uses the university's name and the correct semester.
Running the routine:
- Pause. Two hours feels urgent, but a registration problem can be checked in five minutes.
- Identify the request. It wants a payment through a link.
- Verify separately. The student logs into the official student portal from a bookmark. There is no outstanding balance and no alert.
- Inspect. The link in the text points to a domain that adds an extra word to the university's name.
Result: a scam, identified without any special software. The student reports it to the university's IT security team, which can warn other students.
Quick reference: red flags vs. safe actions
| Red flag | Safe action |
|---|---|
| Urgent deadline or threat | Pause and verify independently |
| Request for secrecy | Tell someone you trust |
| Payment by gift card, crypto or wire | Refuse; real organizations rarely require these |
| Request for a one-time code | Never share it with anyone |
| Link in an unexpected message | Go to the official site or app directly |
| Unfamiliar sender using a familiar name | Check the full address or number |
| Offer you never applied for | Confirm on the company's official site |
| Voice or video asking for money | Call back on a known number |
If you think you have been scammed
Act quickly and without embarrassment. Scams target smart, careful people every day.
- Stop contact with the scammer. Do not send more money to "recover" losses.
- Call your bank or card provider right away to block payments or cards.
- Change passwords for any accounts you shared or that use the same password, and turn on two-factor authentication.
- Save evidence: screenshots, emails, phone numbers, transaction records.
- Report it to the platform, your bank and your national fraud or consumer protection agency.
- Warn people around you, especially if the scammer impersonated you or someone you know.
Key takeaways
- AI makes scams look and sound real, so judge the request, not the polish.
- Use the routine: pause, identify the request, verify through a separate channel, then inspect links and payments.
- Code words, call-backs and official apps beat any detector.
- If something has already happened, contact your bank, change passwords, save evidence and report it quickly.