The Bank of England (BoE), in collaboration with the BIS Innovation Hub in London, has conducted a pioneering experiment using AI to detect fraud in retail payment systems.
According to Finextra, dubbed Project Hertha, the initiative explored whether modern AI tools could identify complex and previously unseen financial crime patterns in real-time transactional data.
Criminals typically act through intricate networks of accounts across various financial institutions, making detection a significant challenge. Payment systems, however, provide a valuable network-wide perspective, and the BoE aimed to leverage this to detect suspicious activity more efficiently. The research simulated a large dataset of 1.8 million synthetic bank accounts and 308 million transactions, using advanced AI models to mimic real-world transaction behaviour.
The findings show that payment system analytics could serve as a “supplementary tool” for banks and payment service providers (PSPs) to enhance their crime detection capabilities. Participating institutions were able to identify 12% more illicit accounts than they could through conventional means. Moreover, when it came to spotting novel and previously unseen crime patterns, the AI-driven analysis delivered a 26% improvement.
However, the BoE acknowledged the limitations of the approach. While the AI model demonstrated potential, it is not a standalone solution. The bank noted that implementing such a system would involve a host of legal, regulatory, and operational challenges that were outside the remit of Project Hertha.
“The results demonstrate promise but also show there are limits to the application and effectiveness of system analytics,” the BoE stated. “It is just one piece of the puzzle. The introduction of a similar solution would also raise complex practical, legal and regulatory issues. Analysing these was beyond the scope of Project Hertha.”
The central bank also stressed the importance of using labelled training data, a robust model feedback loop, and explainable AI algorithms to improve the effectiveness of such solutions. As the financial sector increasingly turns to AI, the findings from Project Hertha provide both a benchmark and a caution for future innovation.
Keep up with all the latest RegTech news here
Copyright © 2025 RegTech Analyst
Copyright © 2018 RegTech Analyst
