The Alan Turing Institute has published its much anticipated report, commissioned by the UK Financial Conduct Authority, focusing on AI challenges, benefits and transparency as they relate to financial services (and more broadly).
While the report perhaps covers familiar ground around the implications of the design and use of AI, it provides an articulate and succinct overview of the technical aspects of AI in three key areas of innovation relevant to the financial services sector (machine learning, non-traditional data and automation) and the consequential challenges in design and deployment.
Critically, the report neither seeks to apply its conceptual framework on AI transparency to individual use cases of the application of AI in financial services, nor comments on the extent to which existing regulatory arrangements or industry practices are appropriate to mitigate concerns of regulators that may arise in this context. These areas are signposted as important areas of further work to be undertaken. In particular, no recommendation is given as to whether any such regulatory changes (to the extent identified as being required) should be specific to AI technologies (or certain AI technologies, or certain AI technologies when applied to specified use cases in the financial services sector) or could instead be folded into existing regulatory frameworks or guidance.
In view of these comments and in lieu of more concrete guidance around regulatory expectations in the context of the use of AI in financial services, firms should make sure to engage with compliance teams and legal counsel at an early stage when embarking upon or outsourcing any AI project. Attention should be paid to the key themes and areas of concern, shown in this diagram, highlighted in the report, with a view to ensuring that all AI innovations being deployed are properly understood and adequately considered at the appropriate levels and at each relevant juncture of the lifecycle of the applicable AI use case.
Finally, it is worth noting that, although the UK is no longer a member of the EU, the European Commission’s proposed AI Regulation has extra-territorial effect. Firms operating across borders will need to take on board its requirements. For more information, see EU proposes new Artificial Intelligence Regulation.