Jump to a section
The short answer

Scams still work the same way: they rush you, ask for secrecy and want money moved. AI just makes the voice, video or message more convincing. So don't try to spot the fake. Check the request instead: hang up and call back on a number you know, agree a family safe phrase, and never move money for someone you've only met online.

If you've already sent money or shared bank details, call your bank or card provider now, using the number on your card or in its official app. In the UK, you can also dial 159. Speed matters more than anything else on this page.

What AI has actually changed

Scams have always relied on the same few tricks: a sense of urgency, a believable story and a request to move money. AI hasn't changed that. It has made the story more believable and quicker to produce.

The FBI has set out how criminals use it. They write convincing messages faster, use translation tools to remove the spelling mistakes that used to give scams away, create realistic photos for fake profiles, clone a relative's voice from a short clip, and make fake videos, including for live video calls.

That sounds alarming, but notice what's missing from that list: a new way to take your money. Every one of these still ends with someone asking you to pay, share details or keep quiet. That's where you can stop it.

The numbers, in proportion

In the US, people reported losing a record $15.9 billion to fraud in 2025, up from $12.5 billion the year before, according to Federal Trade Commission figures reported by the Detroit Free Press. Investment scams made up $7.9 billion of that, nearly half. Impostor scams, where someone pretends to be a person or organisation you trust, were the most common, with more than a million reports.

The FBI's Internet Crime Complaint Center counted $20.9 billion in losses in 2025, AARP reports, with $7.7 billion lost by people aged 60 and over.

Those are big numbers, but the pattern behind them is simple, and that's the good news: almost every one of these scams ends with a request to move money. That's the moment to stop.

In the UK

In the UK, the banking trade body UK Finance found that criminals stole £1.28 billion through payment fraud in 2025, up 4%. Banks also stopped £1.68 billion of unauthorised fraud before it was taken.

The biggest growth was in investment scams, up 40% to £221.5 million. Romance scam losses rose 23% to £39.2 million. Impersonation scams, where someone pretends to be your bank or the police, fell by 12%.

Two-thirds of the UK scams where people were tricked into sending money started online. So the places to be most alert are social media, adverts and messaging apps, wherever you live.

The five AI scams worth knowing

You don't need to know every variation. These five cover most of what's out there, and each has a simple answer.

Common AI scams and what to do
ScamWhat it looks likeWhat to do
Cloned voiceA call from a family member in trouble who needs money nowAsk for your safe phrase, or hang up and call them back
Deepfake celebrity advertA well-known face on video, promising big investment returnsDon't call the number or click through. Scroll on
Polished messageA text or email from your bank, a parcel firm or HMRC, with no spelling mistakesDon't use its links. Go to the official app or website yourself
Fake profileA friendly new contact online, with lovely photos and maybe a video callNever send money to someone you've only met online
Fake officialA video or phone call from 'the police' or 'your bank'Hang up and call back on a number you look up yourself

Use AI to plan Christmas and the holidays: gifts, food, budget and scams

Deepfake adverts: how one case unfolded

Consumer group Which? followed the case of a woman who saw an Instagram video that appeared to show Nigel Farage on Question Time, talking about a way to make millions by investing. It was fake. The video gave a phone number. She called, was passed to a 'broker', and paid £200 by card to start.

The 'broker' used a slick trading website that claimed to be regulated. It wasn't authorised by the Financial Conduct Authority. When she later tried to send £17,000, her bank stopped the payments. Which? reports that the scammer had already warned her the bank would probably disapprove.

That last detail is the giveaway. Anyone who tells you to keep a payment from your bank is a scammer. Which? also names Martin Lewis, Rishi Sunak, Elon Musk and Bear Grylls among the faces used in other deepfake scams.

Which?'s advice is simple: be wary of pressure to act fast, guaranteed or unusually high returns, requests for secrecy, and any unexpected contact about investments. For big money decisions, use an FCA-regulated financial adviser, or free guidance from MoneyHelper or Citizens Advice.

Cloned voices: the safe phrase fix

Starling Bank says criminals can clone a voice from as little as three seconds of audio, the kind of clip many of us have posted online. In its 2024 survey of 3,010 UK adults, 28% said they had been targeted by an AI voice cloning scam at least once in the past year, and 46% had never heard of them.

The fix costs nothing. Agree a safe phrase with your close family: something only you'd know, never written down online or sent by text. If 'your son' calls needing money, ask for it. The FBI recommends the same thing.

Our family plan guide walks you through setting it up in five minutes.

AI voice scams: a simple plan to protect your family

Fake profiles and romance scams

Romance scams have always been about trust built slowly. AI helps scammers keep it up. The FBI says criminals use AI to create realistic profile photos, send 'private' pictures, and even make videos to prove an online contact is a real person.

So a video call is no longer proof. The rule that works is the FBI's: never send money, gift cards, crypto or anything else of value to someone you've only met online or by phone, however long you've been talking.

How to tell if a photo or video is AI-made, and what to do when you can't

Why 'spot the fake' isn't the answer

You'll see lots of advice about spotting deepfakes. The FBI lists clues like distorted hands, odd teeth or eyes, wrong shadows, lag and unnatural movement. They're worth knowing, but treat them as a bonus.

The more reliable test is the request, not the picture. Does it want you to rush? To keep quiet? To move money or share a code? If so, stop and check using a route you choose, however real it looks or sounds.

Forward scam texts to 7726

Here's a handy one that works in both the US and the UK: forward scam texts to 7726 (it spells SPAM on a keypad). The US Federal Trade Commission says it helps your wireless provider spot and block similar messages, and the UK regulator Ofcom says it's free and lets your provider investigate and block the number.

In the UK: dial 159 for your bank

In the UK, 159 connects you safely to your bank. Stop Scams UK says it reaches customers of more than 99% of the UK's retail bank current accounts and can't be spoofed. If you get an unexpected call about your money, hang up and dial 159. Elsewhere, hang up and call the number on the back of your card.

If it's already happened

There's no shame in being caught out. What matters now is acting quickly.

Contact whoever you paid through, straight away, using a number you trust, like the one on your card. The US Federal Trade Commission's advice works anywhere: for a card or bank transfer, call your bank or card issuer and ask them to reverse the payment; for a gift card, contact the company that issued it; for cryptocurrency, contact the exchange or ATM operator.

Helping older parents use AI safely: a practical guide for families

Where to report it

Then report it. Reports help police spot patterns and warn others.

  • United States: report to the FTC at ReportFraud.ftc.gov, and online crime to the FBI's Internet Crime Complaint Center at ic3.gov.
  • Canada: report to the Canadian Anti-Fraud Centre online, or on 1-888-495-8501, and contact your local police.
  • Australia: report to Scamwatch. Scamwatch also points cybercrime victims to a police report through cyber.gov.au, and says to call 000 if anyone is in immediate danger.
  • Anywhere else: your national police or fraud reporting service.

If you're in the UK

In England, Wales and Northern Ireland, report fraud to Report Fraud, which replaced Action Fraud, at reportfraud.police.uk or on 0300 123 2040. In Scotland, call Police Scotland on 101.

Ask your bank about getting your money back. Which? explains that UK victims tricked into sending money can be reimbursed in some cases, with compensation capped at £85,000.

Pass it on

The best protection is a family that talks about this. Share the checklist below, agree your safe phrase, and make it normal to say 'I got a weird call today'. If you're helping an older relative, our guide for families has a gentle way to start.

Helping older parents use AI safely: a practical guide for families

November watch: checked 12 November 2026

A current UK warning worth sharing with family: on 27 October, Report Fraud said it had received more than 1,300 reports of AI-enabled investment fraud covering January to August 2026, with £24.3 million lost. Those figures describe that reported category and period; they are not all UK fraud or losses during November.

The warning describes fake endorsements using well-known people or trusted institutions. The useful response is to check the firm independently using the FCA's official route, rather than deciding whether a celebrity video looks convincing. Do not use contact details supplied by the advert as your independent check.

Five habits that beat AI scams

  • Slow down. Pressure to act now is the clearest sign of a scam
  • Hang up and call back on a number you look up yourself, like the one on your bank card
  • Agree a safe phrase with close family, and never share it by text or online
  • Never send money, gift cards or crypto to someone you've only met online or by phone
  • Treat investment tips in adverts, videos and messages as a red flag, whoever seems to be in them

Sources

These sources support the guide. Tool settings, prices and policies change, so dated source boxes show when those details were checked.

AI helps with research and drafting. Dan checks the sources and is responsible for what is published.

How we check