The Artificial Intelligence Consortium (AIC), co-chaired by David Geale and Sarah Breeden, held its fourth quarterly meeting at the Financial Conduct Authority's (FCA) offices. The consortium aims to foster dialogue on the capabilities, development, deployment, use, and potential risks of artificial intelligence (AI) in UK financial services.
Discussions included challenges in explainability and transparency for generative AI (GenAI). A workshop identified recurring model risk issues such as system complexity, limited transparency from third-party providers, and output variability, which can make it difficult to apply existing model risk frameworks proportionately.
Another workshop analysed how AI adoption might alter contagion pathways across the financial system. Potential impacts include price volatility, changes in participant behaviour, and system-wide disruption during periods of stress. Members also discussed the importance of distinguishing between visible and invisible risks, and the need for cross-firm approaches to testing contagion.
The consortium also considered concentration risk arising from AI providers in UK financial services. A survey suggested that concentration exists at the model and compute levels, with limited alternatives, particularly for services supporting Important Business Services (IBS).