AI Is Making Fraud 4.5 Times More Profitable, FATF Says

Fraudsters are shifting from sprawling scam compounds to smaller teams using artificial intelligence, according to Giles Thomson, president of the Financial Action Task Force (FATF). For PSPs and banks, the practical issue is that AI is making scams easier to scale, harder to spot and more effective during onboarding and impersonation checks.

1. Thomson told the Financial Times on Monday, Sept. 14, that authorities have gotten better at taking down the old model of large scam compounds, but criminals are now moving to setups like “one or two people in a basement with a very big server.”
2. He said criminals — many in Southeast Asia — have long used trust-building schemes before pushing victims into cryptocurrency and investment frauds or romance scams. AI now lowers the cost of that work by letting scammers impersonate people with deepfakes, run chatbots, and create fake websites.
3. Thomson also said AI is being used in money laundering and bank onboarding through deepfakes and clones, which matters because those are exactly the kinds of signals traditional controls are built to catch.
4. The FT cited Interpol data showing AI-enabled fraud was 4.5 times more profitable than traditional methods, with overall fraud-related reports up 54% since 2024.
5. Interpol said agentic AI systems can autonomously plan and execute complete fraud campaigns, from reconnaissance to ransom demands. The FBI also said earlier this year that it received complaints about $893 million of losses from fraud carried out with AI in 2025.

PYMNTS noted earlier this year that AI helps scammers bypass standard fraud detection, which has traditionally depended on spotting anomalies in transaction patterns or behavior during onboarding. Visa’s James Mirfin told PYMNTS in March that in a café or restaurant, people’s behavior is usually fairly similar — which is exactly why those older signals are getting harder to rely on.