Bank Of England Explores AI Solutions To Detect Emerging Threats In Retail Payment Systems

According to reports, the Bank of England has collaborated with the London Centre of the BIS Innovation Hub to explore the use of artificial intelligence (AI). Their aim is to detect new and unusual patterns of financial crime in real-time retail payment systems. Although the project has shown potential, several challenges have emerged that could impact its overall success.

Experts reveal that criminals often attempt to hide their activities by operating through complicated webs of accounts spread across various financial institutions. However, electronic payment systems, which handle transactions among many participants, offer a wide view of these activities across the network.

Bank Of England Aims To Enhance Crime Detection In Payment Systems

Project Hertha, led by the Bank of England, explored whether modern AI tools could help identify such complex and coordinated criminal behaviour within payment system data. The experiments relied on a cutting-edge synthetic dataset specially created for the project, which includes information from 1.8 million bank accounts and covers 308 million transactions.

This extensive dataset was built using an advanced AI model designed to mimic real-world transaction behaviours, enabling realistic analysis. Together, these elements provide a robust foundation for testing the effectiveness of AI tools in detecting financial crime.

The findings suggest that analysing payment system data can act as a useful additional resource for banks and payment service providers (PSPs) in detecting suspicious financial activity. Participants in the project were able to detect 12% more illicit accounts than they typically would have without the use of these AI tools. The approach also proved especially helpful in identifying new or previously unknown forms of financial crime, showing a 26% improvement in spotting these novel behaviours.

Bank Of England Wants To Balance AI Innovation With Regulatory Considerations

The Bank of England states that while the findings show promise, they also highlight the limitations of relying solely on system-level analytics. The central bank emphasizes that this method is only one part of a broader solution. Additionally, deploying a similar system in the real world would raise many practical, legal, and regulatory challenges that fall outside the scope of Project Hertha.

The bank also notes that the effectiveness of such systems depends heavily on the availability of clearly labelled training data, strong feedback mechanisms for the models, and the use of explainable AI algorithms.

The BIS Innovation Hub plays a central role in supporting innovation within the global central banking community. Based within the Bank for International Settlements, it works to identify how technological developments affect central banking. It also aims to create shared tools for the financial system and serves as a platform for experts to collaborate.

The platform operates through an international network of centres and its Innovation Network, which encourages knowledge exchange and coordinated efforts among central banks.