European officials warn AI is moving faster than financial rules

European financial authorities are warning that artificial intelligence is advancing faster than conventional rulemaking, raising questions for market integrity, cyber resilience and financial stability. Officials at the European Central Bank, the Bank of England and the U.K. Financial Conduct Authority said AI could improve productivity, but its speed of development is straining supervisory tools.

Nikhil Rathi, chief executive of the FCA, told CNBC’s “Squawk Box Europe” that the standard regulatory timetable is poorly suited to technologies that can change within weeks or months. He said authorities need new ways to work with markets, including on financial crime and AI-related risks, while avoiding unnecessary barriers to adoption.

“Technology moves incredibly fast, and we need to think differently about some of the innovations that we are seeing on AI,” Rathi said, according to CNBC. He added that regulators should be clear about where risks sit, even as they seek to support the use of new tools.

Regulators focus on agentic AI

The concern centres in part on agentic AI, systems designed to take actions with greater autonomy. In finance, that could include models that assist with research, operations or, over time, trading decisions. Greater autonomy may increase efficiency, but it can also make it harder for firms and supervisors to identify how a decision was made or to halt harmful activity quickly.

Rathi pointed to work in Britain on frontier AI, including the U.K. AI Safety Institute, as part of efforts to give policymakers, companies and regulators a better understanding of the technology’s risks. He also cited the need to protect market integrity as AI becomes more widely embedded in financial services.

Christine Lagarde, president of the European Central Bank, told France’s Les Échos that AI offers potential productivity gains, while also creating a “major risk.” She linked the concern to long-running threats such as cyberattacks, hacking and data theft, and said faster and more advanced AI models pose a more serious challenge because defensive tools and the necessary funding have not yet been established.

Trading safeguards under discussion

AI’s effect on productivity and market integrity was also discussed at the ECB’s annual forum in Sintra, Portugal, this week. Sarah Breeden, deputy governor of the Bank of England, said in a speech on Tuesday that agentic AI could intensify market swings during periods of stress.

Breeden said trading firms currently tend to use autonomous AI for lower-risk operational work, including research, but that this could change quickly. She said wider use in markets may call for safeguards similar to circuit breakers or kill switches, which could limit or stop trading across markets if defective AI models contributed to severe disruption.

The debate comes as European officials also acknowledge that the region trails the United States in AI investment and in the development of leading frontier AI companies. Boris Vujčić, vice-president of the ECB, said Europe needs to develop its own AI capabilities, with sovereignty concerns featuring in the policy discussion. He said Europe has previously adapted new technologies to raise productivity growth, although it has not consistently been at the technological frontier.

According to CNBC, investors have argued that AI-related spending has helped support U.S. outperformance, while Europe’s bank-centred financial system gives the region fewer channels to finance AI investment. That financing gap adds an economic dimension to the regulatory challenge: authorities are trying to encourage adoption while setting controls for risks they say are still developing.

This story draws on original reporting from CNBC.