European banks say AI, crypto make fraud more complex

A survey of 200 European fraud leaders presented at a 2026 webinar hosted with NICE Actimize found banks reporting rising fraud complexity driven by AI-initiated payment flows and scams tied to virtual assets. Respondents said efforts to deploy AI-based controls are limited by integration issues, fragmented data and funding shortfalls.

Panelists on the webinar included Chris Ainsley, head of fraud risk management at Santander UK; Joe Bristow, product director and fraud subject-matter expert at NICE Actimize; and Sharon Kimathi, the session moderator. The discussion covered shifting fraud patterns across channels and how institutions are adjusting controls.

Respondents reported that traditional threats such as card fraud and account takeover remain significant. Newer vectors identified in the survey include automated payment requests generated by AI and scams involving virtual assets. Respondents described these newer schemes as increasing in both frequency and technical sophistication.

Banks are expanding use of artificial intelligence for transaction monitoring, behavioral analytics and cross-channel detection to spot anomalous activity more quickly. Several institutions said they are piloting models that combine signals from payments, account activity and external data sources, including virtual asset transaction records.

Survey participants highlighted practical barriers to broader AI rollout. Integration challenges with legacy systems, inconsistent or fragmented data sets, limited budgets and gaps in specialist staffing were listed as factors that prevent some banks from scaling AI solutions and embedding them into existing risk platforms.

The responses showed uneven preparedness across European markets and firms. Larger banks with dedicated fraud budgets and in-house data science teams are more likely to test and integrate advanced AI tools. Smaller banks and some regional players reported greater reliance on vendor solutions or manual processes while seeking funds and expertise to upgrade defenses.

On confidence in current fraud controls, many teams expressed concern that existing measures will not be sufficient over the next few years without substantial upgrades. Reported investment priorities include improving detection accuracy, cutting false positives and integrating multiple signal types to improve investigators’ ability to trace suspicious flows.

Panelists and respondents noted that effective AI deployment depends on data quality, orchestration across channels, regulatory compliance and ongoing model governance. They said those requirements add workload for fraud, risk and compliance teams already operating with constrained resources.

Background trends cited in the survey include increased use of automation and synthetic identities by fraudsters to scale attacks across payment rails, and the growth of virtual assets as an avenue for illicit transfers and potential money laundering. Respondents said tracing and blocking illicit flows that move outside traditional payment systems remains a challenge for banks.