The Financial Conduct Authority (FCA) has made a significant about-turn in its approach to artificial intelligence within the financial sector with the publication of the Mills Review, a landmark assessment commissioned by the regulator. This shift marks a major departure from the FCA's previous stance, which advocated for applying existing regulatory frameworks to AI without specific rules, as if it were simply an adjunct to human decision-making.
Authored by Omar Salem, architect of the FCA’s Consumer Duty, the Mills Review delves deep into the implications of AI in retail financial services, extending far beyond simple chatbots. A central theme is the rise of 'AI agents' – sophisticated programmes capable of undertaking complex tasks, accessing personal documents, and acting on behalf of individuals with third parties such as banks and wealth managers. These agents have the potential to revolutionise how consumers manage their finances, from comparing savings accounts and switching insurance to monitoring pension contributions and managing investments.
The review acknowledges the significant benefits of AI agents, particularly in addressing long-standing issues within retail finance. It points to the low uptake of traditional financial advice, the considerable number of unbanked individuals, and the vast sums held in low-interest accounts, suggesting AI could empower consumers who are disengaged or poorly served. Indeed, findings indicate that one in five UK adults are already open to AI making financial decisions for them, with around 26 per cent trusting general-purpose AI tools for financial advice.
However, the Mills Review also grapples with critical questions surrounding trust and accountability. Concerns are raised about potential scenarios where an AI agent might mismanage funds, make poor investment choices, or even lead to unfair discrimination through personalised pricing. Previously, the FCA largely left it to firms to interpret how existing regulations applied to AI. Now, the review recommends that the FCA consider mandating specific requirements for AI agents to bolster consumer protection.
Intriguingly, the review also suggests the FCA itself should leverage AI to supervise firms and their AI agents, a move that, while seemingly circular, reflects a commitment to utilising the best available technology for regulatory oversight. For City firms, the message is clear: robust AI governance is no longer optional but a fundamental capability. The era of navigating AI compliance in a regulatory vacuum is drawing to a close, ushering in a period of increased clarity alongside greater scrutiny.
The FCA's shift does not abandon its principles-based regulatory model, which has been a strength of the UK system. Instead, the Mills Review outlines how AI is evolving from mere assistance to active decision-making, compelling the regulator to keep pace with this transformative technology and ensure that consumer interests remain paramount.