FCA publishes Mills Review on AI risks and opportunities

The Financial Conduct Authority has published The Mills Review, a study led by FCA director Sheldon Mills that maps how artificial intelligence could reshape UK retail financial services by 2030 and warns it may amplify risks including fraud, cyber attacks and consumer harm. The review was announced in January and published in July 2026.

The report draws on FCA research estimating that around a fifth of UK adults — about 11 million people — are likely to use AI systems that can act autonomously within pre-set goals. It states wider use of such systems could improve access, personalisation and efficiency for customers while increasing the scale and speed of harms across the sector.

Mills writes that AI will change how firms operate, how consumers make financial choices and how markets function. He identifies four main shifts: firms’ operations and product design, new ways consumers make decisions, altered competitive dynamics and an amplification of fraud and cyber risk. The review highlights consumer concerns about trust and control of AI systems and recommends safeguards to address those concerns.

The report sets out steps for regulators and government, calling for active monitoring of AI systems, clearer expectations for firms that use AI, stronger testing and reporting requirements, and investment in regulatory capabilities. It encourages firms to use FCA testing initiatives to trial AI under oversight and suggests regulators should adopt AI tools to improve supervision and enforcement.

The publication follows criticism from MPs on the Treasury Committee, who in January warned that the Bank of England and the FCA had taken a “wait-and-see” approach that left the public and the financial system exposed to “potential serious harm.” The review presents a roadmap for regulators and government to prepare for projected AI-driven changes and recommends closer collaboration across agencies to manage cross-cutting risks such as cyber security, fraud and market concentration.

Ashley Alder, chair of the FCA, noted the review highlights potential benefits for consumers and firms and outlines how risks can be managed. The report does not propose a single regulatory model but urges timely action and ongoing engagement with industry to ensure rules keep pace with technology.

Separately, the Bank of England has begun exploring operational tools to limit market disruption from AI, including so-called “kill switches” that could halt trading if models behave unpredictably. At the ECB’s Sintra Forum, deputy governor Sarah Breeden said agent-like uses of AI in commerce and trading raise questions about whether existing, technology-agnostic regulatory frameworks remain sufficient.