UK retail banks and building societies are accelerating the deployment of artificial intelligence systems to detect and prevent fraud, with industry-wide investment in AI-driven security tools expected to exceed £900 million in 2026, according to new figures from UK Finance.
The surge in spending comes as authorised push payment fraud and identity theft continue to rise, with criminals increasingly exploiting sophisticated social engineering techniques to bypass traditional security measures. Banks reported a 12 per cent increase in attempted fraud incidents during 2025, though the proportion of cases resulting in financial loss fell slightly, suggesting that detection systems are becoming more effective.
“The fraud landscape has changed fundamentally,” said Margaret Okonkwo, head of financial crime prevention at the Banking Standards Board. “We are no longer dealing with crude phishing emails. Today’s fraudsters use AI-generated voice cloning, deepfake video calls, and highly personalised scam narratives built from data harvested through breaches. The only way to fight AI-enabled crime is with AI-enabled defence.”
The technology being deployed ranges from real-time transaction monitoring that analyses thousands of data points per second to behavioural biometrics that can identify users by the way they type, swipe, and navigate banking apps. Several major institutions are also piloting systems that use natural language processing to scan live chat and phone conversations for indicators of coercion or deception.
Barclays, Lloyds Banking Group, and NatWest have each announced multi-year partnerships with technology firms to build proprietary fraud detection platforms. Smaller challenger banks are taking a different route, licensing modular solutions from specialist providers such as Featurespace and ComplyAdvantage, which can be integrated without the need for extensive in-house data science teams.
The regulatory environment is also evolving. The Financial Conduct Authority has signalled that it expects firms to treat AI fraud detection as part of their consumer duty obligations, and proposed new rules would require banks to reimburse fraud victims in most circumstances unless they can demonstrate that adequate detection systems were in place.
Privacy advocates have raised concerns about the scope of data collection involved in behavioural monitoring, and the Information Commissioner’s Office is consulting on updated guidance for financial services AI. Industry leaders say the balance between security and privacy is achievable with proper governance.
“No one is suggesting we surveil customers without consent,” Okonkwo said. “But when a customer’s life savings are on the line, most people are willing to accept a degree of automated monitoring if it keeps their money safe. The key is transparency about what data is collected and why.”