AI in financial services is creating new opportunities for banks to improve areas such as fraud detection, anti-money laundering compliance, and risk management. Machine learning systems can analyze large volumes of data and identify patterns that might be difficult or time-consuming for employees to detect, making the technology increasingly attractive to financial institutions.
However, greater reliance on artificial intelligence also creates new governance challenges. UK regulators are examining how banks and other financial firms use AI and machine learning, with particular attention to issues such as biased decision-making, accountability, and management oversight. As adoption increases, financial institutions will need to ensure that automated systems operate within clearly defined ethical and regulatory boundaries.
The executive director for UK deposit-taking supervision executed a surveyed more than 200 banks, insurers and financial market infrastructure firms regarding the use of AI for detecting money-laundering with results to be release at the end of the year. In the meantime banks better be on the lookout for biased AI implementations:
‘Are data being used unfairly to exclude individuals or groups, or to promote unjustifiably privileged access for others?’ Proudman said, adding that recent examples of retailers using overly-personalised marketing can seem plain ‘creepy’.
Boards will also need to consider how to allocate individual responsibilities under the Senior Managers Regime, which requires every activity at a bank to come under the direct responsibility of a named official so it is easier for regulators to identify and punish them when things go wrong.
‘You cannot tell a machine to ‘do the right thing’ without somehow first telling it what ‘right’ is – nor can a machine be a whistle-blower of its own learning algorithm,’ Proudman said.
‘As the rate of introduction of AI/ML in financial services looks set to increase, so too does the extent of execution risk that boards will need to oversee and mitigate.’
The potential benefits of AI in financial services are significant, particularly in areas such as anti-money laundering where institutions must continuously analyze enormous volumes of transactions and customer information. However, automating those processes does not eliminate the responsibility financial institutions have for the decisions their systems make.
Bias represents one important concern. Algorithms trained on inappropriate or incomplete data could unintentionally disadvantage certain individuals or groups, while highly personalized use of customer information could create privacy concerns or undermine consumer trust. Banks therefore need to understand not only what their AI systems can accomplish, but also how those systems arrive at decisions.
Accountability presents another challenge. Under regulatory frameworks that assign responsibility for banking activities to specific executives, institutions cannot simply transfer responsibility to an algorithm. Boards and senior managers will need appropriate oversight mechanisms to understand how AI systems operate, identify potential problems, and intervene when necessary.
As adoption of AI in financial services accelerates, governance will need to develop alongside the technology. Effective implementation will require financial institutions to combine the analytical capabilities of artificial intelligence with human judgment, clearly defined responsibilities, and controls designed to identify unintended consequences before they create significant regulatory or customer risks.
Overview by Tim Sloane, VP, Payments Innovation at Mercator Advisory Group
