The UK Financial Conduct Authority (FCA) has published a landmark 147-page report on artificial intelligence in financial services, known as the Mills Review, after Sheldon Mills, the FCA's Executive Director of Competition. Touted as the first review of its kind by a global regulator, the report warns of a rapid shift from episodic human-led financial decisions to continuous, delegated services driven by agentic AI.
Key findings highlight that by 2030, AI could fundamentally change how firms operate, consumers make decisions, markets compete, and risks emerge. While AI may reduce friction and tackle issues like advice gaps and low switching, it also amplifies fraud, cybersecurity risks, and consumer harm. A striking 20% of UK adults—roughly 11 million people—are already open to letting AI make autonomous financial choices.
The report sets out seven broad recommendations, including the creation of an agentic supervisory model, foundations for 'agentic finance', trusted agent protocols, and expansion of the FCA's AI Lab. Crucially, the review brings programmable money into the regulatory frame. It notes that multi-day settlement in traditional banking rails is too slow for AI agents executing portfolio and transaction strategies, and that systemic stablecoins and tokenized assets on programmable ledgers could provide the instant, atomic settlement needed for machine-directed finance.
However, the FCA insists that human accountability must survive automation. The report warns that liability becomes murky when autonomous systems move money or cause harm, with one CEO even suggesting a 'Turing test' for finance. Ashley Alder, Chair of the FCA, emphasized the need for the UK to keep pace with AI benefits, while Monica Eaton, CEO of Chargebacks911, cautioned that the race between good and bad actors will only accelerate. The review is seen as a blueprint for redesigning financial markets, treating automation not as a software upgrade but as a market redesign that could reshape the role of tokenized assets and stablecoins in everyday finance.
