European policymakers say financial regulation is struggling to keep up with the rapid development of artificial intelligence and how to support its adoption while limiting risks to market health and stability.
Nikhil Rati, chief executive of the UK Financial Conduct Authority, said the traditional rule-making cycle “doesn’t work”, especially in an era of technological change that accelerates the development of agent AI.
“Technology is moving incredibly fast and we need to think differently about some of the innovations we’re seeing in AI,” Rati told CNBC’s “Squawk Box Europe” on Thursday.
Rati highlighted the UK Financial Stability Board’s work on Frontier AI and the establishment of the UK Institute for AI Safety as part of a broader effort to help policymakers, regulators and businesses better understand risks and safely deploy technology.
European Central Bank President Christine Lagarde said AI is a source of productivity and profit. But in an interview with France’s Les Echos newspaper, she also warned that the technology also poses “huge risks”.
“For about 10 years, we’ve been talking about cybersecurity risks, hacking, data theft, etc.,” Lagarde said. “But as AI models accelerate and deepen, we face more serious risks, because it is happening so quickly, and because we have not yet found the means of protection and the funding needed for it.”
Her comments came after the impact of AI on productivity and market health emerged as a key issue at the ECB’s annual general meeting (the European version of the Jackson Hole Symposium) in Sintra, Portugal, this week.
Sarah Breeden, deputy governor of the Bank of England, said agent-based AI could amplify volatility in times of market stress.
Breeden said in a speech in Sintra on Tuesday that for now, trading companies are primarily using autonomous AI for low-risk business tasks such as research. “But that could change quickly,” she said.
What about guardrails and circuit breakers?
He said the increased use of agent AI in financial markets could require tighter oversight, including guardrails “similar to circuit breakers or kill switches” that “restrict or halt market-wide trading if a flawed AI model causes a market meltdown.”
But top bankers and regulators also acknowledge that Europe lags behind in AI investment and the development of frontier companies that drive breakthroughs.
Boris Vujicic, deputy president of the European Central Bank, said: “Europe is currently in a situation where it has to develop its own capabilities in the field of AI, of course. There has also been a lot of talk about sovereignty issues in the field of AI. Europe has shown in the past that it can adapt new technologies … (to improve productivity). (But) it has not always been at the forefront.”
Rati said market authorities ultimately need to strike a better balance when it comes to such rapidly evolving technologies.
He said that while technological innovation offered exciting opportunities for the UK, particularly around productivity and growth challenges, it was important to ensure the market was not exposed to risks that regulators were not yet able to adequately monitor.
“The reality is that some of these technologies are moving in weeks or months, and traditional rule-making cycles don’t work that way. So we need to think about new tools and different ways to work with markets in a more collaborative way on things like financial crime and AI risks to ensure that we can ensure the goal of market integrity.”
He added: “We don’t want to get in the way of adoptions, but we need to be transparent about where the risks are.”
