Artificial intelligence can create realistic voices, images, videos and written messages. While this technology has many useful purposes, criminals can also use it to make scams more convincing.
A caller may sound exactly like someone you know. An email may copy the style of a trusted company, while a video may appear to show a celebrity or business leader promoting a product.
These are often known as AI scams. Some use “deepfakes”, which are fake or altered recordings designed to make it appear that a person has said or done something they have not.
The Financial Conduct Authority has warned that AI could increase the risk of fraud involving deepfakes, fake identities and manipulated communications.
Knowing the warning signs can help you stop, check and avoid being rushed into a decision.
What is a deepfake?
A deepfake is an image, video or audio recording that has been created or changed using AI.
A deepfake could make it appear that:
- A family member is calling you for help
- A celebrity is promoting an investment
- A manager is asking an employee to make a payment
- A company representative is requesting personal information
- A public figure is making an announcement
- Someone is attending a video call when they are not
Deepfakes can sometimes include visible or audible mistakes, but the technology is becoming more convincing. A recording should not be trusted simply because the face or voice appears familiar.
How do AI voice scams work?
AI can be used to copy the voice of a real person. Criminals may use recordings from social media, videos, voice notes or other online content to create fake speech.
You could receive a call or voice message that appears to be from:
- A family member who claims to be in an emergency
- Your bank or insurance broker
- A police officer or government department
- Your employer or manager
- A company you already use
The caller may know some personal information about you. This does not prove that they are genuine, as details may have been found online or taken from an earlier data breach.
Stop! Think Fraud warns that criminals can use AI voice cloning to impersonate someone the victim knows or trusts, often while asking for an urgent payment.
Warning signs of an AI voice scam
Be suspicious if the caller:
- Creates an urgent or emotional situation
- Asks you to send money immediately
- Requests a one-time passcode, password or PIN
- Tells you not to speak to anyone else
- Asks you to move money to a “safe account”
- Wants payment through gift cards or cryptocurrency
- Asks for remote access to your computer or phone
- Refuses to let you call back
- Contacts you from an unexpected number
Do not rely on the number shown on your screen. Scammers can make a call appear to come from a genuine company or person.
How to check whether a call is genuine
End the call and contact the person or organisation using a number you already trust.
If the caller claims to be a relative, call them directly. You could also contact another family member who may know where they are.
Some families choose a private word or question to use during an emergency. It should be something that cannot be found on social media.
If someone claims to be calling about your insurance, use the contact details on your policy documents or your insurance broker’s official website. Do not use a telephone number provided by the caller.
How is AI used in scam emails?
AI can help criminals write clear and professional emails. This means spelling mistakes and poor grammar are no longer reliable signs of a scam.
A fraudulent email could appear to come from:
- An insurance provider or broker
- A bank
- A delivery company
- HM Revenue and Customs
- A streaming or shopping service
- Your employer
- A friend or relative
The message may copy real branding, colours and writing styles. It might even include personal details to make it appear genuine.
Stop! Think Fraud says AI is making phishing emails more convincing, although unusual addresses, urgent instructions, unknown links and requests for personal information remain important warning signs.
Warning signs of an AI-generated scam email
Check for:
- An email address that is slightly different from the genuine one
- Unexpected requests to update your payment details
- Claims that your policy or account will be cancelled immediately
- Links asking you to sign in or confirm personal information
- Unexpected attachments or QR codes
- Requests for passwords or security codes
- Pressure to act before a short deadline
- Changes to bank details for an expected payment
- A greeting or tone that feels unusual
- Website addresses containing extra letters, numbers or misspellings
An email can still be fraudulent even if it uses your correct name and looks professional.
How to check a suspicious email
Do not reply, open an attachment or click a link.
Instead, visit the organisation’s website by entering its address yourself. You can then sign in through the official website or contact the company using a trusted telephone number.
Check the full sender address, not just the name shown at the top of the email. A scammer may display the name of a genuine company while using an unrelated email address.
Suspicious emails can be forwarded to the National Cyber Security Centre at report@phishing.gov.uk.
How do deepfake video scams work?
Deepfake videos are often shared through social media adverts, messaging apps, fake news stories or video calls.
A video may appear to show a celebrity, financial expert or company director promoting:
- An investment opportunity
- A competition or giveaway
- A new financial product
- A cryptocurrency scheme
- A way to recover money lost through fraud
- A limited-time offer
The person shown in the video may have no connection to the product.
Criminals can also use deepfake videos during live or recorded calls. They may pretend to be someone with the authority to request a payment or ask for confidential information.
Warning signs of a deepfake video
Some deepfakes may show:
- Speech that does not match the person’s mouth
- Unnatural facial movements or blinking
- Blurring around the face, hair or mouth
- Changes in lighting or skin tone
- Jewellery, glasses or facial features that flicker
- A flat or unnatural voice
- Sudden changes in sound quality
- Unusual pauses or repeated phrases
However, visual mistakes are not always present. Government fraud guidance warns that AI-generated material can be difficult to identify by sight alone.
The message and behaviour surrounding the video can provide stronger warning signs than its appearance.
Be cautious if the video promotes an unexpected opportunity, promises guaranteed returns or pressures you to act quickly.
How to check whether a video is real
Look for the same information on the person or organisation’s official website and verified social media accounts.
You could also:
- Search for reliable news reports about the announcement
- Check where the video was first posted
- Look for longer or higher-quality versions
- Avoid using links included with the video
- Contact the organisation directly
- Report suspicious adverts to the platform
A familiar face is not proof that an offer is genuine.
Could scammers pretend to be an insurance company?
Criminals may use AI-generated messages or calls to impersonate a genuine financial firm.
An insurance-related scam could claim that:
- Your policy is about to be cancelled
- You are owed an unexpected refund
- A payment has failed
- Your bank details must be updated
- A claim cannot continue until you pay a fee
- More documents or security details are needed
- Someone can recover money for an upfront payment
Fraudsters may copy a company’s name, logo, website or Financial Conduct Authority reference number. The FCA calls these “clone firms”.
Always check that contact details match those on the FCA Firm Checker and the company’s official website. Do not use contact information supplied in a suspicious message.
What tactics do AI scammers use?
The technology may be new, but many of the tactics are familiar.
Scammers often try to create:
Urgency
You are told that something bad will happen unless you act now.
Fear
The message may claim that your money, policy, account or family member is at risk.
Authority
The scammer pretends to be a police officer, manager, financial firm or government department.
Secrecy
You may be told not to discuss the situation with anyone else.
Excitement
A fake video or advert may promise a prize, refund or easy financial return.
Trust
The scammer uses a familiar voice, face, logo or piece of personal information.
Whatever the message looks or sounds like, stop if you are being asked for money, security information or immediate action.
What should you do if you suspect an AI scam?
Follow three simple steps:
Stop
Do not send money, share information, open attachments or click links.
Check
Contact the person or organisation separately using details you know are genuine.
Report
Reporting suspicious contact can help protect other people.
You can:
- Forward suspicious emails to report@phishing.gov.uk
- Forward suspicious text messages to 7726
- Report scam adverts through the platform where they appeared
- Report fraud to Report Fraud in England, Wales or Northern Ireland
- Contact Police Scotland on 101 if you are in Scotland
Ofcom advises people not to engage with suspected scam calls or messages and not to share personal or financial details.
What if you have already responded?
Do not feel embarrassed. AI scams are designed to be convincing, and anyone can be targeted.
Act as soon as possible:
- Contact your bank immediately if you sent money or shared payment details.
- Change any password you revealed, including on other accounts where you used the same password.
- Turn on two-step verification where available.
- Contact your insurance broker if you shared policy or claim information.
- Run a security check if you downloaded a file or allowed access to your device.
- Save copies of messages, email addresses and payment details as evidence.
- Report what happened to the appropriate fraud service.
The quicker you respond, the better the chance of limiting further harm.
Stay cautious, even when something looks real
AI means that seeing a familiar face, hearing a trusted voice or receiving a professional email is no longer enough to prove that contact is genuine.
Focus on what you are being asked to do. Unexpected requests for money, passwords, passcodes or personal information should always be checked through a separate and trusted channel.
If you receive a suspicious message about your insurance, end the contact and speak to your insurance broker using the details shown on its official website or your policy documents. One Call Insurance can help you check the correct next step without requiring you to respond to an uncertain call, email or video.

