AI scams and deepfakes: how to spot them and stop them

What has AI changed?

Scams are nothing new. What's new is that AI lets a scammer produce messages that look and sound genuine, quickly and cheaply. In the past you could often spot a scam by its clumsy language. Today the text can be flawless and address you by name.

AI can also imitate a person's voice from just a few seconds of audio – taken, for example, from a social media video. The same technology is used to make deepfakes: videos in which a well-known person appears to say something they never said.

The core of a scam, though, is always the same: urgency, emotion and money. Once you recognise that pattern, you'll recognise most scams, whether AI was involved or not.

The "emergency call" with a cloned voice

The phone rings. It's your grandchild, in tears: they've had a car accident or been arrested and need money right now. Then a "police officer" or "lawyer" comes on the line and explains how to pay. The voice sounds exactly right – but it's an AI-made copy.

Warning signs:

• The caller insists you act immediately and gives you no time to think.
• You're asked to keep it secret from the rest of the family.
• Money is requested in odd ways: gift cards, cryptocurrency, cash handed to a "courier", or a transfer to an unfamiliar account.
• The number is unknown – or it looks familiar but the call feels off. Phone numbers can be faked.

What to do: hang up and call back yourself on your relative's usual number. If they don't answer, call another family member. Real police will never ask you to pay fines or bail over the phone.

Fake investment ads featuring celebrities

On social media and in news-site ad slots you'll often see "news stories" in which a well-known TV presenter, athlete or business leader says they got rich on a new investment platform. Sometimes there's an AI-made video in which the person seems to speak directly to you. It is almost always a scam, and the celebrity knows nothing about it.

A typical sequence: you enter your contact details, and soon a friendly "investment adviser" calls. Your first small deposit seems to grow nicely. Then you're asked for more money, to install remote-access software, or for your banking codes. The money never comes back.

Warning signs:

• Big, guaranteed returns are promised quickly. Real investing comes with no guarantees.
• The ad mimics a well-known newspaper's design, but the address bar shows a strange web address.
• You're encouraged to install a remote-access program "so they can help".
• The company doesn't appear among authorised firms on the financial supervisor's website – in Finland, Finanssivalvonta (FIN-FSA) – or it appears on their warning list.

Phishing: messages that look real

In phishing, the scammer tries to get you to click a link and enter your login details on a fake page. The message may appear to come from your bank, the postal service, the tax office or a parcel company. Thanks to AI, the language is now fluent, so it no longer gives the scam away.

Check these things:

1. Were you expecting it? If you haven't ordered a parcel, a "delivery problem" is suspicious.
2. Are you being rushed? "Your account will be closed today" is a classic pressure tactic.
3. Where does the link really go? Don't click. Instead, go to the service yourself by typing the familiar address or opening your own app.
4. Are you asked for codes or card details? Authorities and banks don't ask for these via a link in a message.

In Finland, mobile operators already block some spoofed numbers and text messages, but not all of them. Staying alert is still essential.

Can AI help you spot scams?

Yes – as long as you remember it isn't infallible. You can copy the text of a suspicious message into, say, /en/tools/chatgpt/ or /en/tools/gemini/ and ask for an assessment. Remove your own personal details first, and don't click any links in the message.

If the AI says the message looks genuine, still check for yourself on the service's own website or by calling its official number. To check where a picture came from, try /en/tools/google-lens/: it shows where the same image has appeared before.

Spotting a deepfake

Fakes are getting better all the time, so you can't rely on your eyes alone. Sometimes small details give them away: lips not quite in sync with speech, odd blinking, unnatural lighting, or a voice that's too flat. More important than examining details is asking:

• Did this come from the official organisation's own channel?
• Are reliable news outlets reporting the same thing?
• Am I being asked to act right away or to pay?

If it has already happened

Shame is a scammer's best friend, but there's no reason to feel it. Scammers are professionals, and anyone can be fooled. The faster you act, the better your chances of limiting the damage.

1. Call your bank immediately and ask them to block your cards and online banking codes. Banks' blocking services work around the clock; the number is on the bank's website and often on the back of your card. Ask them to stop any payments if it's still possible.
2. Change your passwords for any service whose login you gave away. Start with your email.
3. If you gave remote access to your computer, disconnect it from the internet and get IT help before you use it for banking again.
4. Report it to the police. In Finland you can file a report online on the police website or at a police station. Keep messages, call logs and transfer details as evidence.
5. Report it to the National Cyber Security Centre – in Finland, the NCSC-FI at Traficom. Reports help warn others about the same scam.
6. Tell your family. Scammers often try again, for instance by offering to "recover your money" for a fee. That's a scam too.

If something is happening right now and someone is in danger, call the emergency number (112 in Finland and across the EU).

A short checklist for the whole family

• Agree on a safe word and keep it within the family.
• Urgency and secrecy are always a red flag.
• Hang up and call back on a number you know.
• Don't click the link in a message – go to the service yourself.
• Never give codes, card details or remote access to anyone who contacted you first.
• Limit public videos and voice clips on social media where you can.

Talk about this especially with older relatives. A calm conversation in advance is the best protection there is. More tips in /en/guides/ai-for-seniors/ and /en/guides/ai-privacy-and-security/.