Report underscores the need to understand and manage the risk implications of AI in banking

• The application of AI to financial services is bringing benefits both to customers and to banks
• But, as with any new technology, it carries risk implications for individual banks and the financial system
• Regulators need to create frameworks for the adoption of AI while industry needs to work with them to apply new technology to regulation itself

The power of Artificial Intelligence (AI) to improve productivity in financial services is undeniable but it also brings new risks, according to a report published today. The report, “Sustainable AI in Finance: Understanding the Promises and Perils” by Parker Fitzgerald, a risk management consultancy, says the technology is benefiting customers with faster decision-making and a more dimensional view of customers and markets. At the same time, predictive analytics and machine learning have opened new possibilities in the detection of fraudulent activity and financial crime.

For example, report contributor Ayasdi, a pioneer in the application of AI to financial services, is engaged with HSBC to improve its Anti-Money Laundering (AML) systems by developing an intelligent segmentation of customers. This fine-grained understanding of customer behaviors reduced the number of false positives by 20% while enhancing the overall risk profile for the bank.

Other financial institutions are using AI to change the ways they optimise capital, model risks, and manage their businesses. For instance, the equity derivatives arm of one bank is using unsupervised learning algorithms to detect anomalous projections generated by its stress testing models.

This technology promises to transform the industry, but it also brings with it new risks. Until now the overwhelming focus has been on the threats to job security, but the more pressing concern is the financial stability implications of AI. One such example is the new and unexpected forms of interconnectedness between financial markets. Additionally, the use of AI may also make it difficult for human users at financial institutions – and for regulators – to grasp how decisions, such as those for trading and investment, have been formulated.

As the applications of AI continue to grow, the report sets out three principles for managing the risk:

1. regulators need to specify their “red lines” for the use of AI by companies. Explainability auditability, and reproducibility will be key in governing the use of AI and other technologies in finance.
2. greater RegTech use will be critical for improving regulatory efficiency. Further use of ‘tried and tested’ tools, like the FCA’s FinTech Sandbox could also prove effective.
3. macro-level standards on AI and international data regulations will be integral to the responsible adoption of AI. This is pertinent in the context of Brexit.

Sustainable AI in finance requires a deeper understanding of the technology in view of its risks. This will require a holistic approach that encompasses data, technology, process, and governance.