European banks bolster AI controls as AI and crypto fraud rise

A survey of 200 European fraud leaders, conducted in association with NICE Actimize and presented during a webinar moderated by researcher Sharon Kimathi, maps the fraud landscape for 2026.

Respondents reported a sharp rise in AI-initiated payment schemes and fraud tied to virtual assets. The survey found fraudsters increasingly use automation and generative tools to create payment instructions, impersonate customers and move funds into digital wallets.

Traditional card and online banking fraud remain active while risks grow in mobile payments, push-payment scams and transfers involving virtual assets. The spread of activity across channels has increased the complexity of detection and response for fraud teams.

Banks reported investing in machine learning models, behavioural analytics and rules engines to flag suspicious flows generated or assisted by AI. Firms also described efforts to strengthen controls around virtual asset custody and on- and off-ramps for crypto transactions.

Respondents identified obstacles to rapid modernisation, including difficulties integrating new AI tools with legacy infrastructure, challenges accessing and cleaning the data needed for reliable models, and a shortage of staff with the required skills. Funding and capability varied across markets, with larger banks more likely to deploy advanced solutions than smaller institutions.

The survey highlighted budget priorities such as AI model deployment, data integration projects and hiring or reskilling fraud specialists. Banks are also prioritising cross-channel monitoring so alerts from card, account and virtual asset systems can be correlated more quickly.

Chris Ainsley of Santander UK and Joe Bristow of NICE Actimize discussed how combining rule-based systems with machine learning can cut false positives and how sharing anonymised fraud patterns across industry networks can speed detection of emerging tactics.

Background material presented during the session noted that fraud patterns differ by market because of payment habits, regulatory regimes and technology adoption, and described fraud as becoming faster, more automated and more cross-channel.