AI risks · Fraud and scams

How do scammers use AI?

Criminals already use AI to write convincing messages, clone voices and fake videos of real people. The scams themselves are old; AI makes them cheaper, more polished and harder to spot.

Happening nowSocietal impact:

How AI is used in fraud

In December 2024 the FBI described how criminals use generative AI in financial fraud. Text tools produce phishing messages without the spelling mistakes that used to give them away, and help run romance and investment scams. Image tools create fake profiles, forged ID documents and fake celebrity endorsements. Voice cloning is used to imitate a family member in a staged emergency, and real-time video is used to pose as executives or police officers.

The International AI Safety Report 2026 confirms documented cases where scammers used voice clones and deepfakes to impersonate executives or relatives and trick victims into transferring money. It also notes that the tools are cheap, require no expertise and can be used anonymously.

The Swedish picture

Sweden's Financial Supervisory Authority (Finansinspektionen, FI) launched an information campaign on AI fraud in March 2026. Its survey shows that six in ten Swedes are worried about AI being used for fraud, with concern somewhat more common among women and older people. FI describes two typical patterns: a cloned voice that sounds like a relative urgently asking for money, and a manipulated video where a well-known person appears to recommend an investment. These are often combined with professional-looking websites, fake customer reviews and time pressure.

Why the numbers are uncertain

It is hard to know how much fraud AI actually adds. Victims often do not report out of embarrassment, and police and bank statistics rarely record whether AI was involved. The International AI Safety Report describes systematic data on the scale of AI-enabled fraud as limited.

What is being done

Authorities focus on awareness and simple routines, since the weak point is usually the moment of pressure rather than the technology. The EU AI Act's labelling requirements for deepfakes and AI-generated content, which apply from August 2026, aim to make synthetic media easier to recognise, though criminals are unlikely to follow them. AI companies and researchers are also working on detection, but detection tools still lag behind generation quality.

What you can do

  1. Agree on a code word or a control question with your family to use if someone calls asking for money in an emergency.
  2. Hang up and call back on a number you look up yourself, even if the voice or face seems familiar.
  3. Never share codes, login details or card information with someone who contacts you, however credible they seem.
  4. Treat investment ads featuring famous people with suspicion, and check the firm in the financial regulator's register before paying anything.

Sources

  1. Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud (PSA I-120324-PSA) – FBI Internet Crime Complaint Center (IC3), 2024
  2. FI varnar för AI-bedrägerier – så skyddar du dig – Finansinspektionen, 2026
  3. Skydda dig mot AI-bedrägerier – Finansinspektionen, 2026
  4. International AI Safety Report 2026 – International AI Safety Report (chair: Yoshua Bengio), 2026

Data updated: 2026-11-19

More risks

AI risks