FCA Plans AI-Driven Supervision, No New Rules for Chatbots

The UK Financial Conduct Authority (FCA) plans to automate as much of its supervision and enforcement as possible using artificial intelligence, its data chief told lawmakers on Tuesday. The regulator is not preparing new AI rules for the industry, according to the Financial Times.

Jessica Rusu, the FCA's chief data, information and intelligence officer, said AI works well on the emails, documents and interview records that firms provide. She added that the FCA is deploying AI in ways that keep its officials overseeing and directing the work, and is looking at embedding it in authorizations, supervision and enforcement. The regulator wants to go "further than that, to automate wherever we can . . . and to intervene at scale," she said.

Rusu is due to set out the FCA's formal answer to the Mills Review in a speech next week. Former FCA executive Sheldon Mills delivered the review in July with seven priority recommendations, including building an AI-enabled supervisory model. The review also said the regulator would need greater powers to keep up with AI and urged UK authorities to consider whether large language models such as ChatGPT, Claude and Gemini should fall under their rules, according to the FT.

Nikhil Rathi, Chief Executive of the FCA, told the House of Commons Treasury Committee on Tuesday that his thinking on AI had been evolving. He is still not convinced the FCA needs "lots more powers" until it has shown it has used its existing powers to the fullest extent. A model that gives information without an individual recommendation that results in a transaction does not fall fully inside the regulated perimeter, he said, applying to advice, insurance and other markets. "I think that is going to be a very live discussion," Rathi said, noting the open point is where to draw the line.

FCA staff have visited firms to see how they use AI and will publish their findings on good and poor practice next week, Rusu said. The AI Live Testing service opened in December 2025 with NatWest, Monzo, Santander and Scottish Widows in the first cohort. The testing follows a sandbox that gave banks NVIDIA computing power to trial AI models in June 2025. Five months later the FCA agreed to test AI tools jointly with the Monetary Authority of Singapore. The FCA's first technology horizon scan in June named AI as one of three shifts likely to reshape financial services by 2030.

Outside the regulatory perimeter, research commissioned for the Mills Review found that a fifth of UK adults are open to letting AI models make financial decisions for them, with savings and borrowing among the examples. Meanwhile, at least ten brokers and platform vendors connected AI agents to live client accounts in the first half of 2026, and Anthropic's Claude powered nine of them, an FM Intelligence study found. eToro began letting investors hand trades to their own AI agents in March, with the agents running in sub-accounts with a set budget and scoped access. In June, ThinkMarkets launched ChelseaAI, an MCP server that lets a trader's AI assistant place orders but can block it from client funds. In each case the authorized firm is the broker holding the account, not the developer of the model.