AI Scams: 12 Warning Signs and How to Protect Yourself
November 6, 2026
Not long ago, a scam email was easy to spot: broken grammar, blurry logos, ridiculous requests. Those days are over. With AI scams, fraudsters now write flawless messages, mimic the tone of your bank or your boss, clone a loved one’s voice from a few seconds of audio, and even show up on a video call wearing a deepfake face.
The good news: AI changes the disguise, not the script. Every scam, however polished, needs three things: urgency, a channel the scammer controls, and an action you can’t undo (a payment, a code, a password). In this guide, we cover the 12 warning signs of AI scams, the defenses that actually work, and what to do right away if you’ve already clicked or paid.
Why AI scams are more dangerous
Generative AI has made tailor-made fraud cheap. A personalized attack used to take hours of work. Now a scammer can:
- write thousands of flawless emails and texts, in any language, tailored to each target;
- gather public information (social profiles, job sites, reviews) to make a message believable;
- clone a voice from a public video or a voice note;
- generate synthetic faces and video for fake video calls, fake endorsements, or “guaranteed” investments backed by celebrities;
- run entire conversations with chatbots that reply around the clock with endless patience.
The result: old advice like “look for spelling mistakes” is no longer enough. You need a method.
12 warning signs of an AI scam
1. Urgency is the first red flag
“Your account will be locked within 24 hours.” “Pay today or your package will be returned.” “I need a transfer right now.” Pressure is designed to stop you from thinking. No real bank, courier or government agency expects you to decide in a few minutes.
2. The message is too perfect (and too personal)
A polished message that mentions your name, your city, your latest purchase or your manager’s name is not proof of authenticity. That information is often available online, and AI can piece it together in seconds.
3. They ask you to switch channels
An email that tells you to “continue on WhatsApp,” a text that makes you call a number, a social media message that moves you to Telegram. Pulling you onto a channel they control is one of the most common moves in the book.
4. The sender looks right, but the domain doesn’t
Check the full email address, not just the display name. [email protected] is not your bank. Watch for domains with a swapped letter, an extra hyphen or an unusual ending.
5. The link leads to a login page
If a message sends you straight to a page asking for your username, password or codes, stop. Open the official website or app yourself, by typing the address or using a bookmark, and check from there whether anything actually needs your attention. This is especially important with package delivery and bank texts, a scam known as “smishing.”
6. They ask for a code you just received
A one-time passcode sent by text or app exists to confirm your action. Anyone who asks you to read it out, even if they claim to be your bank or “tech support,” is trying to get into your account. No legitimate company will ask for it.
7. The voice is familiar, but the request isn’t
“Mum, I lost my phone, this is my new number, I need money.” With AI, a short clip is enough to imitate someone’s voice. If the request is unusual, a familiar voice proves nothing.
8. Something is off on the video call
Real-time deepfakes often show lips out of sync, a fixed stare, face edges that “shimmer” when the person moves, or inconsistent lighting. Ask the person to turn their head to the side or wave a hand in front of their face. Fakes often glitch or distort.
9. The investment is “guaranteed” by a famous face
Videos of entrepreneurs, journalists or celebrities pushing trading platforms or crypto with “guaranteed returns” are among the most common AI-powered scams. No legitimate investment guarantees profits, and public figures don’t pitch them in random ads.
10. They want payment in a hard-to-recover form
Instant bank transfers, cryptocurrency, gift cards and prepaid top-ups are scammers’ favorites because once the money is sent, getting it back is very difficult. Anyone who asks you to pay a bill, fine or fee with gift cards is running a scam.
11. They tell you to keep it secret
“Don’t tell anyone.” “This is a confidential transaction.” “Don’t mention it to your bank.” Secrecy isolates you from the people who might spot the fraud. It’s a very strong warning sign, especially at work.
12. The story falls apart under a simple question
Ask something only the real person would know: a shared memory, a family detail, something that isn’t on social media. Scammers, whether human or chatbot, tend to dodge the question or pile on more urgency.
Quick table: real or fake?
| Situation | Legitimate behavior | Scam signal |
|---|---|---|
| Bank | Points you to the official app | Texts a link to “unlock” your account |
| Delivery company | Tracking on the official website | Asks for a small fee to “release” your package |
| Family member | Is happy for you to call back | Insists on a new number and an immediate transfer |
| Boss or supplier | Follows company procedures | Demands an urgent, confidential payment |
| Tech support | Never asks for passwords or codes | Wants your one-time code or remote access to your computer |
7 defenses that actually work
- Always call back yourself. Hang up and call the person or company on a number you already know or find on the official website, never the one in the message.
- Set up a family safe word. Agree on a secret word or question to use in an emergency request. It’s the simplest and most effective defense against cloned voices.
- Turn on two-factor authentication for email, banking, social media and messaging apps. Where possible, use passkeys: they can’t be handed over to a fake site.
- Use a password manager. It fills in credentials only on the correct domain. If it doesn’t autofill on a page, that’s a strong clue the site is fake.
- Limit what you share publicly. Fewer public voice notes, videos and personal details means less material for cloning your voice or building a convincing story.
- At work, use a “two-channel” rule. Any payment request or change of a supplier’s bank details must be confirmed on a second, independent channel, even if it appears to come from the boss.
- Talk to the people most at risk. Parents, grandparents and teenagers are favorite targets. Ten minutes of conversation is worth more than any antivirus.
Already clicked or paid? What to do right now
- Contact your bank immediately on the number on the back of your card or in the official app, and ask them to block cards and stop any pending payments. In the first few minutes, a transfer can sometimes still be stopped.
- Change your passwords for every affected account, starting with your email, and sign out of all active sessions.
- Turn on or reset two-factor authentication, and check that no unknown devices, automatic forwarding rules or recovery numbers have been added.
- Keep the evidence: screenshots, phone numbers, email addresses, receipts. Don’t delete the conversation.
- Report it. Contact your local police. In the US, report it to the FTC at reportfraud.ftc.gov. In the UK, forward scam emails to [email protected] and scam texts to 7726. If your account was used to scam others, warn your contacts.
Frequently asked questions
Can AI really clone my voice?
Yes. With just a few seconds of clear audio, for example from a public video or a voice note, there are tools that can produce a very similar voice. That’s why a familiar voice on the phone isn’t enough to confirm a request for money.
How can I tell if an email was written by AI?
Often you can’t, and that’s the point: what matters isn’t who wrote it, but what it asks you to do. Urgency, links to login pages, and requests for codes or payments are scam signals no matter how good the writing is.
Does antivirus software protect me from AI scams?
It helps block known malicious sites and attachments, but it can’t stop a phone call with a cloned voice or a transfer you make yourself. Your main protection is a habit: stop, verify on another channel, and never act under pressure.
What should I do if I get a suspicious video call from a relative?
Hang up and call back on the number you know. If you’ve agreed on a family safe word, ask for it. A real relative won’t mind an extra check.
The bottom line
AI scams are more convincing, but they always follow the same pattern: urgency, a change of channel, and an irreversible request. Slow down, verify through a channel you control, and never share codes or passwords. Save this guide and share it with the people you care about. It’s the simplest way to protect them too.