Don't fall for an AI scam: why scammers are more convincing than ever

The rapid advances in artificial intelligence (AI) tools have armed scammers with powerful new methods to steal money and personal information – and they can be frighteningly convincing.

Here, we reveal some of the most common AI scams, explain how you can spot them, and share ways to protect yourself.

What’s on this page?

• What is an AI scam?
• AI powered romance scams
• Deepfake videos
• Voice-cloning scams
• AI-written phishing emails and texts
• AI-generated scam websites
• How AI is fighting back
• Stay vigilant: 8 red flags that could signal an AI scam
• Reporting scams

What is an AI scam?

An AI scam uses artificial intelligence (AI) to create convincing fake messages, images, videos or even voices. Fraudsters rely on technology that can generate realistic content from just a few words or short audio clips, making it easier to trick you.

These tools are freely available online and have many genuine uses. But their accessibility has led to a surge in scams that are becoming harder to detect. Simon Miller, director of policy, strategy and communications at fraud prevention service Cifas, says: “AI allows criminals to collect data quickly, generate convincing fake documents in seconds, and use deepfake technology to impersonate trusted individuals.

“Older savers are particularly attractive to fraudsters because a single successful hit can get them years of savings.” You might receive a call that sounds exactly like your grandchild urgently asking for money, for example, or see a convincing video of a well-known celebrity endorsing a fake investment.

AI powered romance scams

Scammers are using AI to create fake dating profiles and emotionally manipulative messages. These romance scams often start on dating websites or social media platforms, where fraudsters pose as genuine people looking for companionship. The goal is to build trust and then ask for money.

AI helps criminals write realistic messages and generate fake voice and video calls. Barclays has seen a 20% rise in romance scams in the first three months of 2025 compared to the same period last year. Victims lost £8,000 on average last year, but those aged 61 or over are most at risk, with average claims totalling £19,000.

Kirsty Adams, fraud and scams expert at Barclays, says: “Falling for a romance scam can be both financially and emotionally devastating. We need tech companies to step up their support and take greater action to prevent scammers from operating on their platforms.”

Be cautious of anyone who seems too good to be true, moves the relationship forward quickly, or avoids meeting in person, she warns. If someone asks for money, even for an emergency, speak to a friend, relative, or your bank before sending anything. Stick to messaging through the original platform rather than switching to email or WhatsApp, as this offers an extra layer of security.

Deepfake videos

AI tools can now produce realistic videos that show someone appearing to say something they haven’t. These so-called ‘deepfakes’ are often used to promote fake investments or impersonate public figures. A recent example mimicked well-known financial expert Martin Lewis. The fake video used a cloned version of him to promote an investment scam on social media.

Viewers were invited to join a WhatsApp group for investment tips. This wasn’t the first time the presenter had been cloned, and he’s previously warned that people without more regulation, people will continue to be scammed.

Jenny Ross, editor of Which? Money, says: “The rapid rise of AI offers a whole new avenue for fraudsters to dupe people, and many deepfake videos can be unnervingly accurate. But there are tell-tale signs. Watch the video carefully for lip-syncing that's slightly off.”

Other signs include strange head movements or repeated gestures. Be wary of robotic-sounding voices, unnatural blinking, or other body movements that seem odd.

Voice cloning scams

It’s now possible for scammers to clone someone’s voice using just a short audio clip. This could be a voicemail, video, or even a social media post. They might then call pretending to be a loved one, claiming there’s an emergency and asking you to send money. Because the voice sounds familiar, people often act before thinking.

Alternatively, they may pretend to be from your bank or another trusted authority. Ross says: “There are things you can do to test whether they are a real person. AI doesn’t yet react well to laughter, so laughing or saying something off-topic could trip them up.”

“You should also ask them for detail or a particular memory if they’re a family member, because only a real caller is likely to be able to answer accurately.” Setting up a family code word or shared question can also offer a simple way to check someone is who they say they are.

Try asking unexpected questions or referring to personal details only a real person would know. If they hesitate or dodge the question, that’s a warning sign.

AI-written phishing emails and texts

Phishing emails and texts are nothing new, but AI has made them far more convincing. Scam messages used to be littered with spelling mistakes and odd grammar. Now, AI tools can produce slick, professional-looking emails that appear to be from trusted organisations like banks, HMRC, or even your energy provider.

They might use your name, include the correct logos, and even reference genuine account details to make it look authentic. But there are still signs to watch out for such as vague greetings, unfamiliar senders, messages urging you to act quickly, or links that don’t look quite right. Always double-check the sender’s address or phone number.

If it looks suspicious, don’t click any links or download attachments. Call the company directly using the number on their official website to confirm the message is genuine.

AI-generated scam websites

Some scammers are now using AI to build realistic fake websites that look almost identical to genuine ones. These sites may purport to be banks, pension providers, or even government services.

These scam sites often ask you to enter personal details, make a payment, or log in with your personal details. Always check the web address carefully. Look for small changes like misspellings or unusual endings. If in doubt, search for the official website separately and contact them directly to confirm.

How AI is being used to fight back

While criminals are using AI to commit fraud, technology is also helping to fight back. Banks, for example, use AI to detect unusual transactions and block suspicious payments before they go through. Mobile phone providers are also using AI to filter out scam texts and calls before they reach your device.

Email services are becoming smarter at flagging phishing attempts and stopping them before they land in your inbox. O2 has even launched an AI chatbot called d’AIsy (Daisy) that talks to scam callers to waste their time, stopping them from targeting others.

The Pensions Regulator is also using machine learning to detect and remove fake websites aimed at pension savers, preventing fraud before it happens.

Stay vigilant: 8 red flags that could signal an AI scam

Keep your money safe by looking out for these warning signs. And trust your instincts. If something feels off, it probably is. It’s always better to be cautious.

1. Unexpected contact: Scammers often use AI to automate phone calls, emails, and even WhatsApp messages that look genuine. If someone contacts you out of the blue claiming to be from your bank, HMRC, or a service provider, be cautious.

2. Requests to switch platforms: Fraudsters may ask you to move a conversation from a trusted platform like Facebook or LinkedIn to WhatsApp or email. This is to avoid detection. Always keep conversations on the original platform.

3. Pressure tactics: Scammers thrive on panic. They might say your money is at risk or that you need to act quickly to secure an investment. Genuine companies will never rush you.

4. Highly convincing fake websites: With AI, scammers can now build websites that look almost identical to the real thing. Check the web address carefully for misspellings or extra characters and avoid clicking links from unsolicited emails. 

5. Demands for personal information or payments: If someone asks for your bank details, password, or payment for a service you didn’t request, it’s a major red flag. 

6. Voice cloning attempts: If you get a call from a relative or friend asking for money in an emergency, call them back directly on their known number to confirm. 

7. Deepfake videos or endorsements: Be wary of online videos of celebrities or public figures promoting investments. These can be deepfakes designed to look real.   

8. Generic greetings: While AI has made scam emails more polished, some still slip through with vague greetings like ‘Dear customer’. Trust your instincts if it doesn’t look right. 

Reporting scams

If you suspect you’ve been targeted, contact your bank immediately. Report the scam to Action Fraud (0300 123 2040) or Police Scotland if you’re in Scotland. You can also speak to your local Citizens Advice or Victim Support for support.                                                 

If you’ve shared personal details, change your passwords and run a virus scan on your computer. Forward scam texts to 7726 and report suspicious emails to report@phishing.gov.uk. Fake websites can be reported to the National Cyber Security Centre.

You might feel embarrassed about having fallen victim to a scam. But remember, these scams are designed to be convincing, and anyone can be taken in. Reporting helps protect others and means you can get support.