The Financial Conduct Authority (FCA) plans to use artificial intelligence to automate much of its supervision and enforcement work, while holding off on introducing new AI-specific rules for financial services firms.

Jessica Rusu, the FCA's chief data, information and intelligence officer, told MPs on Tuesday that the regulator wants to expand its use of AI to analyse emails, documents and interview records submitted by regulated firms. The approach forms part of plans to develop an AI-enabled supervisory model that allows officials to identify potential risks and intervene at scale.

Rusu said the FCA wants to go “further than that, to automate wherever we can ... and to intervene at scale”. The regulator is exploring applications across authorisations, supervision and enforcement, with human officials retaining oversight of decisions.

Chief executive Nikhil Rathi told the House of Commons Treasury Committee that the FCA is not yet convinced it needs significantly greater powers to regulate AI. Rathi said he wanted the watchdog to demonstrate that it had used its existing powers to their fullest extent before seeking more.

The debate extends to whether general-purpose AI models such as ChatGPT, Claude and Gemini should fall within financial services regulation. Rathi said chatbots providing general information, rather than personalised recommendations that lead to transactions, do not currently fall fully within the regulated perimeter.

“I think that is going to be a very live discussion,” Rathi told MPs, referring to where the regulatory boundary should be drawn. The question follows the July Mills Review, led by former FCA executive Sheldon Mills, which recommended an AI-enabled supervisory model and called for consideration of whether large language models should be subject to financial rules.

Research commissioned for the review found that one-fifth of UK adults would be open to allowing AI models to make financial decisions for them, including decisions involving savings and borrowing, according to the Financial Times.

The FCA is already testing AI applications. Its AI Live Testing service opened in December 2025, with NatWest, Monzo, Santander and Scottish Widows among the first participants. The regulator has also worked with the Monetary Authority of Singapore on testing AI tools and previously provided banks with access to NVIDIA computing resources through a sandbox initiative.

Rusu said the FCA would publish findings on good and poor AI practices at financial services firms the following week, alongside its response to the growing use of advanced AI models. The regulator's approach currently relies on existing frameworks, including Consumer Duty, rather than a separate set of AI-specific requirements.

---