FCA signals regulatory shift as AI moves to scale in UK financial services

The U.K. financial regulator is signaling that artificial intelligence in finance has moved from experimentation to deployment at scale, and that its own approach will shift with it. In a speech on Wednesday, Financial Conduct Authority Chief Executive Nikhil Rathi said the regulator is rethinking how it supervises firms as AI becomes mainstream, with a sharper focus on trust, competition, fraud and operational resilience.

The message matters for banks, asset managers and fintechs because it offers a clear policy direction from one of the world’s most important financial watchdogs. This was not a new FCA rulebook or enforcement action. It was a speech pointing to how the regulator intends to encourage AI deployment while paying closer attention to broader market structure, consumer harm and system-wide risks.

Rathi, speaking at techUK’s “Agents of Change: Generative and Agentic AI in Financial Services 2026” on June 24, said the FCA is “re-thinking what it means to be an effective regulator in the age of AI,” with emphasis on trust, competition and resilience. The text published on the FCA website is labeled a drafted speech and may differ from the delivered version. He said “more than 80% of financial services firms are already adopting AI,” arguing that the key question is no longer initial uptake but whether the U.K. becomes the place where AI is used safely at scale. “We’re building the future together, and nowhere is that clearer than on AI,” Rathi said. Independent research from the Cambridge Centre for Alternative Finance at Cambridge Judge Business School found about 81% of surveyed financial firms report some AI adoption, supporting his point that uptake is already widespread.

Rathi pointed to two main scaling opportunities: agentic systems and tokenization. Agentic AI refers to systems that can coordinate actions or carry out transactions with limited human input. Tokenization refers to programmable, on-chain financial infrastructure. In the speech, Rathi said: “on Monday, we approved Baillie Gifford, alongside Bank of New York Mellon, to launch the U.K.’s first natively tokenised authorised fund.” He also pointed to the FCA and Bank of England’s joint Call for Input on tokenization in U.K. wholesale markets, published in May 2026.

He said the FCA is also using AI itself. The regulator is “exploring agentic AI as our ‘first responder’ to speed up how we monitor wholesale markets,” Rathi said, adding that the FCA processes about 1 billion rows of data per day. He highlighted support programs for firms including the Supercharged Sandbox, which he said includes partners such as NVIDIA and soon Google, as well as the FCA AI Lab, a new “Agentic Academy,” and an AI Consortium with the Bank of England.

Rathi paired that push with warnings about harm and fragility. He cited UK Finance data showing almost 1.3 billion pounds, or about $1.8 billion, was lost to payment fraud in 2025, and said two-thirds of authorized fraud cases begin on online or social platforms. More broadly, financial services is a critical sector for AI because it provides funding, infrastructure and trust, but firms are also becoming more dependent on third-party cloud providers and AI model providers, raising resilience and concentration concerns already on regulators’ radar.

His message was that the FCA wants supervised experimentation, not a hands-off approach. “It’s whether we are the place where AI is scaled. In a way that is safe, responsible and commercially viable,” Rathi said. He also signaled the FCA expects to use its system-wide powers more often where AI creates wider market or competition problems.