In the report’s foreword, Mills wrote, “The central shift is from human-led, episodic financial activity towards services that are AI-enabled, continuous and delegated.”
Published after the FCA opened a review into advanced AI in January, the report presents seven recommendations for future policy. Among them are creating trusted protocols for agentic finance and expanding the regulator’s AI Lab to help financial firms test AI models in a controlled environment.
Agentic AI increases demand for programmable money
Rather than focusing only on today’s chatbot technology, the FCA describes the rapid emergence of “agentic AI,” where software can independently carry out financial tasks across an autonomy spectrum. At the highest level, the report says, humans become observers while AI systems continuously manage financial decisions.
According to the FCA, more than 20 frontier AI models have been introduced since late 2025, accelerating the development of autonomous financial services far faster than previous regulatory expectations anticipated.
Mills also warned that financial institutions are moving beyond recommendation engines. “Firms are moving from systems that recommend actions to systems empowered and trained to take them, and consumers will soon gain agents that act on their behalf,” he wrote.
Research cited by the FCA found that one in five UK adults would already consider allowing AI to make financial decisions autonomously.
As these AI systems become capable of executing multiple transactions without human approval, the report notes that traditional banking infrastructure may struggle to support machine-speed financial activity. Because stablecoins and tokenized bank deposits operate on programmable distributed ledger networks, they can settle transactions instantly through automated execution instead of relying on conventional multi-day settlement processes.
The report therefore identifies programmable forms of digital money as infrastructure that could support autonomous financial services if adoption continues.
Human accountability remains central despite automation
Alongside the technological opportunities, the FCA dedicates significant attention to governance and legal responsibility. It warns that firms cannot delegate accountability to algorithms even if AI systems execute financial decisions independently.
Industry participants consulted during the review highlighted growing uncertainty over legal liability. According to the report, one chief executive suggested that financial markets could eventually require a “Turing test” to distinguish between genuine human decisions and autonomous algorithmic activity.
Commenting on the publication, Emma Banymandhub, chief executive of The Payments Association, said the review reinforces the need for firms to address governance before autonomous AI becomes commonplace.
“The FCA’s Mills Review reinforces that firms should treat agentic AI as an accountability and governance issue now, while providing greater confidence to innovate responsibly as AI adoption accelerates.”
She added that AI offers significant opportunities for financial services, but its long-term success depends on clear accountability, sound governance, and maintaining consumer trust.
Ahead of the report’s release, Mills also told the Financial Times that responsibility cannot be transferred to software. “You need a human on the hook for what they’re doing,” he said, underlining the FCA’s position that management remains accountable even as financial services become increasingly automated.
crypto.news
