UK financial services AI regulation: FCA Mills Review calls for expanded powers

The UK Financial Conduct Authority (FCA) has recommended a formal review into regulating general-purpose AI in retail financial services, signaling a shift in how AI tools like ChatGPT, Claude, and Gemini will be governed when used for consumer finance decisions. The Mills Review, published by FCA executive director Sheldon Mills, calls for securing and adapting the regulatory perimeter within six months to cover AI-driven personal finance management across savings, investments, pensions, mortgages, and debt management. For AI builders, this means tighter oversight of AI products used in financial advice and a need to prepare for expanded governance requirements.

What happened

The FCA-commissioned Mills Review found that one in five UK adults are open to using general-purpose AI tools for financial decisions, particularly around debt advice, pensions, and investments. At the same time, about 75% of UK-based financial services firms are already adopting AI, with insurers and large banks leading the charge. The review warns that consumers may not be aware they have no formal recourse if an AI tool gives bad advice, and that risks include fraud, scams, market disruption, and regulatory arbitrage.

The review's key recommendation is to secure and adapt the regulatory perimeter within six months to explicitly cover AI-driven personal finance management. This includes examining how consumers use AI for savings, investments, pensions, mortgages, and debt management, and assessing implications for competition, innovation, and growth. The FCA also calls for stronger powers to oversee AI adoption, governance, accountability, and consumer protection, and for ongoing monitoring of frontier model capabilities and AI agents as they take more autonomous actions on behalf of consumers.

Why AI builders should care

If you are building AI products for financial services, the Mills Review signals that the regulatory environment is about to become more defined and potentially more restrictive. The FCA wants to move beyond its current principles-based approach (relying on Consumer Duty and Senior Managers Regime) toward a more explicit framework for AI. This means your product's governance, accountability structures, and consumer protection mechanisms will face greater scrutiny.

The review specifically calls out frontier models like ChatGPT, Claude, and Gemini as targets for potential direct regulation. If your product uses these models for financial advice, you may need to demonstrate compliance with new rules around transparency, recourse, and harm monitoring. The six-month timeline for perimeter review means changes could come quickly.

Practical implications

For AI builders and fintechs, the immediate action is to audit how your AI products interact with retail financial decisions. If your tool provides advice or recommendations on debt, pensions, investments, mortgages, or savings, it may fall within the expanded regulatory perimeter. You should:

• Review your governance framework for AI-driven financial advice.
• Ensure clear accountability for AI decisions, as the FCA emphasizes the Senior Managers Regime.
• Prepare for potential intervention tools and mechanisms from the FCA, including the ability to test AI use with the regulator.
• Monitor the upcoming review and engage with the FCA's consultation process.

The review also recommends that the FCA monitor for evidence of consumer harm and new business models, so early compliance could be a competitive advantage.

Caveats

The regulatory landscape for UK financial services AI regulation is still evolving. The Mills Review is a recommendation, not a final rule. The six-month review timeline and specific powers are subject to government and parliamentary action. Additionally, the evidence in this article is based on the Mills Review report and supporting sources; actual regulatory outcomes may differ. AI builders should treat this as a strong signal of direction rather than