FCA shifts from study to real-time supervision of AI in finance, urging clearer disclaimers

The UK’s Financial Conduct Authority is shifting from studying artificial intelligence in finance to supervising it in real time, saying it will immediately fold six of seven recommendations from its July Mills Review into ongoing work, publish examples of good and poor AI practice at firms, and press AI platforms to make the limits of AI-generated financial advice clearer to consumers.

Jessica Rusu, the FCA’s chief data, information and intelligence officer, set out that approach in a speech, “Building the future of finance in the age of AI,” published by the regulator on Monday. The FCA also published a same-day paper, “AI in financial services: good practice and areas for improvement,” drawing on its testing and supervisory work to show where firms are performing well and where they are falling short.

Rusu said the regulator’s response to the Mills Review, published July 6, would not be another consultation. “Our focus is testing, assurance, monitoring and surveillance: directly with industry to understand how AI systems are being deployed,” she said.

In practice, that means more direct engagement with firms through the FCA’s innovation services, including its AI Lab, as the regulator examines how companies are using AI in products, operations and decision-making. Rusu also said AI is no longer only a retail-finance issue. It is reshaping wholesale markets as well, including how firms analyze information, trade, monitor activity and make decisions.

The FCA also tied its AI work to cyber risk. Rusu said the regulator is working directly with banks to understand how frontier AI models are changing cyber defense, including by identifying vulnerabilities and helping with triage, patching and recovery. Those lessons are feeding into the FCA’s supervisory and policy response, she said.

A large part of the watchdog’s concern is now consumer protection. Rusu said more consumers are turning to general-purpose AI tools such as ChatGPT, Gemini and Claude for financial information and advice, while regulated firms are beginning to integrate services into those platforms. That could widen access to financial tools, she said, but it could also blur the line between regulated and unregulated activity and the protections consumers receive.

The speech cited FCA consumer research showing that four in five less-experienced investors ages 18 to 40 have used AI to help with investing. But only 40% of consumers correctly recognize that there is no recourse if something goes wrong after acting on advice from a general-purpose AI tool.

Rusu said the FCA’s answer is not to expand its remit but to make the existing boundary clearer. “So we are calling on AI providers to help consumers understand the limits of financial information and advice generated by their models through clear, proportionate and consistent disclaimers,” she said.

The regulator said it would not create a new standalone public-interest financial capability tool, as the Mills Review had suggested. Instead, it plans to build on existing services, including InvestSmart and the FCA Firm Checker.

The Mills Review, led by FCA executive director Sheldon Mills, examined how AI could reshape retail financial services by 2030 and laid out seven priority recommendations for the watchdog. Monday’s speech signals that the FCA is moving away from broad review work and toward hands-on testing of systems already being used across finance.

That approach also fits with the wider UK regulatory stance. In May, the Bank of England, the FCA and HM Treasury said frontier AI models could materially change cyber risks for financial firms, while emphasizing that companies should manage those risks under existing operational-resilience rules rather than wait for a new AI-specific rulebook.