Bank of England signals regulatory shift as agentic AI poses new risks to financial stability

Agentic AI may require regulatory reform, BOE's Breeden says

LONDON, June 30 (Reuters) - More sophisticated regulatory frameworks may be needed to monitor and contain the risks AI poses to the financial system, one of the Bank of England's deputy governors said on Tuesday. Speaking at the European Central Bank Forum on central ⁠banking in Portugal, Sarah Breeden, deputy governor for financial stability, said the rapid rise in capabilities of AI agents that can act autonomously had exposed potential gaps. "Our frameworks were not built to contemplate autonomous agents, and relying on a human in the loop for all agent actions is unlikely to ⁠be realistic. More sophisticated governance and accountability frameworks may be needed," Breeden said. Regulators and global standard-setting bodies have repeatedly warned about the risks posed by the rollout ⁠of AI across the financial sector since Anthropic released Mythos that could pose significant cybersecurity challenges to the banking industry, ⁠analysts say. The Financial Stability Board earlier in June called for tighter safeguards to guard against ⁠the risks of AI agents, which, it said, posed a distinct challenge to human oversight. Reporting by Phoebe Seers; Editing by Sharon Singleton and Barbara Lewis Our Standards: The Thomson Reuters Trust Principles., opens new tab

Put a 'kill switch' in AI traders to stop meltdown, says Bank of England

Companies using artificial intelligence systems for autonomous trading could be required to install a "kill switch" to avoid mayhem in markets if they go wrong, a senior Bank of England official has said. In a sign that regulators are becoming uneasy about their current hands-off approach to AI, BoE deputy governor Sarah Breeden said it was examining "whether guardrails are needed" for the technology's use in autonomous trading in financial markets. "AI is reshaping finance at speed," Breeden told the European Central Bank's annual conference in Sintra, Portugal. "It's on us to ensure that the next technology surprise does not become a test of financial stability." AI-powered trading could make markets more volatile in periods of stress, Breeden said, because of an increase in "herding behaviour" as autonomous agents "respond similarly to the same prompts or triggers". The BoE was working with Germany's Bundesbank and the Basel-based Bank for International Settlements to examine this herding problem, she said. They were exploring "mitigants" -- including whether AI systems could be given public policy objectives and "whether guardrails are needed, analogous to circuit breakers or kill switches that would limit or stop trading marketwide if faulty AI models cause market meltdown". However, IMF director for monetary and capital markets Tobias Adrian warned the conference: "Kill switches and circuit breakers may not work that well in private markets and in less liquid and over-the-counter markets." Itay Goldstein, a professor at the University of Pennsylvania, argued that AI-based trading activity posed "quite serious challenges" for law enforcement agencies. "When regulators are trying to prevent market collusion in a world with humans, the idea was always to look for signs of communication, co-ordination [and] intention," he told the conference. "With AI you will not find any of this," said Goldstein, arguing that AI trading agents "just learn how to do it" but do not communicate with each other. He pointed to experimental research showing that AI-based trading agents over time "tend to find ways to achieve a collusive outcome" and limit aggressive competition to maximise long-term profits. Comparing AI models to mischievous teenagers, Breeden said: "They lie, they tell you they've not done things when they have and they behave differently when you're watching them." Regulators would need to identify a human who is "accountable for that model", she added. While trading firms were still mostly using autonomous AI models for "lower-risk operational tasks" such as research, Breeden said, "that could change quickly". "The financial system is likely to evolve into one that operates more autonomously, at scale and speed," she said. "The transition is uncertain and will bring risks of its own to monitor." While algorithmic trading has existed for many years, Breeden's comments indicate that regulators are growing more concerned about the potential for the latest AI models to create a new wave of computer-driven activity in financial markets. Regulators have been reluctant to write specific rules for AI in financial services, saying any regulation would soon be out of date as the technology is moving so fast and they fear stifling innovation in a critical area for economic growth. But Breeden said the growing use of AI to power trading and other areas of financial services such as retail payments raised questions about this approach. Once AI-powered payment agents could book a holiday, refill a fridge or refresh a wardrobe, Breeden said, "the biggest issues are likely to be for regulation and industry standards". She said rules may need adapting for "how users give consent and authorisation to agents", as well as for multiple transactions, dispute resolution, fraud and protocols for interacting with merchants and financial institutions. "What these two examples -- agentic commerce and agentic trading -- both highlight is that, as AI capabilities increase, we must keep asking whether existing, technology-agnostic regulatory frameworks remain sufficient," she said.

Bank of England's Breeden warns AI agents could trigger market meltdowns

Deputy governor Sarah Breeden says autonomous trading agents could amplify volatility if they all react the same way at once, and may demand new rules. The nightmare a central banker describes is rarely a crash. It is a feedback loop. Speaking at the European Central Bank's annual forum in Sintra, Portugal, the Bank of England's deputy governor Sarah Breeden warned that autonomous artificial intelligence agents could cause a "market meltdown." Not by acting irrationally, but by acting identically, all of them responding the same way to the same signal at the same moment. Breeden's concern is specific to a new generation of AI. The worry is not the algorithmic trading that has driven markets for years, but agentic systems that can pursue goals and make decisions with far less human supervision. If many firms deploy agents trained in similar ways on similar data, she argued, those agents could "amplify volatility in stress," reacting to a shock in lockstep and turning a wobble into a rout before any human has time to intervene. The mechanism is the danger. Markets have always been vulnerable to herd behaviour, but human herds are slow and uneven; people hesitate, disagree, and panic at different speeds. A population of AI agents optimised toward the same objectives could move as one, selling into the same decline or chasing the same trade with a synchronised speed and scale no crowd of traders could match. The result would be sharper swings, faster, with the correlation between agents acting as an accelerant. Breeden, who speaks for the Bank on financial stability, suggested the existing rulebook may not be equal to the problem. More sophisticated regulatory frameworks may be needed to monitor and contain the risks AI poses to the financial system, she indicated, a notable signal from a senior policymaker that the tools built to oversee human-run markets might not capture what happens when the participants are autonomous software. The warning is not isolated. The Bank of England has been flagging AI-related risks to financial stability for months, folding them into its broader assessments of what could go wrong in markets, and Breeden's Sintra remarks sharpen a theme regulators on both sides of the Atlantic have begun to circle. The question of who is accountable when an autonomous agent acts is becoming a live one for finance, not a hypothetical. It echoes anxieties being raised well beyond the trading floor. The challenge of governing AI agents that act with minimal oversight has produced a wave of work on how to give them verifiable identities and control, precisely so their actions can be traced and constrained. Breeden's warning is the financial-stability version of the same problem: an agent that can act faster than a human can supervise is useful right up until the moment it is dangerous. There is a tension at the heart of the issue that Breeden's framing acknowledges. The same agents that could amplify a crisis are being adopted because they make markets more efficient, executing faster and cheaper than people can. Regulators are not trying to ban them; they are trying to work out how to keep their benefits without inheriting a system that can unravel at machine speed. That is a harder problem than prohibition, and it is the one the Bank is signalling it intends to take on. Breeden did not propose specific rules, and her remarks were a warning and a prompt rather than a policy. What they establish is that the prospect of correlated AI agents destabilising markets has moved from the seminar room to the speeches of the people who would have to manage the fallout. The Bank has put the question on the table. The answer, including whether the current framework can be stretched to fit or needs rebuilding, is the work that now follows.

BoE calls for bespoke AI regulation

This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community. In a speech at a central banking event in Portugal, deputy governor Sarah Breeden spoke of potential gaps in supervision that have arisen from technology developments and specifically agentic AI when used for payments and trading. "Our frameworks were not built to contemplate autonomous ⁠agents, and relying on a human in the loop for all agent actions is unlikely to be realistic," said Breeden. It is not just the use of agentic AI that is of concern but also the prospect or more severe and more frequent cyber attacks. "We need policy frameworks, internal skills and institutional structures ready for more frequent technology surprises. Cyber risks, agentic trading and agentic payments and commerce are the applications on my radar today. But that list could look different in a year, if not sooner," added Breeden. Her proposals include greater use of scenario analysis and using AI to boost monitoring and risk management, including the creation of digital twins to simulate interactions between banks. Breeden also said that the BoE is considering the use of "enhanced recovery" for core systems in which one bank can take over another's processes during a disruption. In addition, Breeden also referfenced the potential use of kill switches and circuit breakers "that would limit or stop trading market-wide if faulty AI models cause market meltdown". The deputy governor's speech comes days after the Financial Stability Board called for more stringent safeguards for the use of AI agents.

Bank of England's Breeden signals new rules to govern agentic AI

Following years of insistence that existing frameworks were sufficient to mitigate AI risks, Deputy Governor Sarah Breeden said rapid developments in areas like agentic payments and trading had exposed potential gaps that could require a more sophisticated regulatory response. Agentic AI can make decisions and operate autonomously. The Bank of England on Tuesday signalled the need for bespoke AI regulation to contain risks to the ⁠financial system ⁠posed by increasingly capable agentic systems, in a potential shift in its approach to overseeing the technology. Following years of insistence that existing frameworks were sufficient to mitigate AI risks, Deputy Governor Sarah Breeden said rapid developments in areas like agentic payments and trading had exposed potential gaps that could require a more sophisticated regulatory response. Agentic AI can make decisions and operate autonomously. "Our frameworks were not built to contemplate autonomous agents, and relying ⁠on a human in the loop for all agent actions is unlikely to be realistic," Breeden told the European Central Bank Forum on central banking in Portugal. Enhanced recovery and kill switches Breeden said the BoE is considering whether banks need "enhanced recovery" for core systems, allowing one bank to take over another's basic functions during a disruption. Other measures under consideration include fresh guardrails and circuit breakers or kill switches "that would limit or stop trading market-wide if faulty AI models cause market meltdown." According to a Cambridge Centre for Alternative Finance survey, 52% of finance firms are already using agentic AI. In commerce, agents are typically used to recommend products and in trading, firms mostly use autonomous AI for lower-risk operational tasks, though that could change quickly, Breeden said. "If AI agents respond similarly to the same ⁠prompts or triggers, they could amplify volatility in stress - especially if their objectives drift from original goals or public policy objectives." Regulators and global standard-setting bodies have repeatedly warned about the risks posed by the rollout of AI across the financial sector since Anthropic released Mythos, a model that analysts say could introduce significant cybersecurity challenges to the banking industry. The Financial Stability Board earlier in June called for tighter safeguards to guard against the risks of AI agents, which, it said, posed a distinct challenge to human oversight.

Bank of England's Breeden signals new rules to govern agentic AI

LONDON, June 30 (Reuters) - The Bank of England on Tuesday signalled the need for bespoke AI regulation to contain risks to the financial system posed by increasingly capable agentic systems, in a potential shift in its approach to overseeing the technology. Following years of insistence that existing frameworks were sufficient to mitigate AI risks, Deputy Governor Sarah Breeden said rapid developments in areas like agentic payments and trading had exposed potential gaps that could require a more sophisticated regulatory response. Agentic AI can make decisions and operate autonomously. "Our frameworks were not built to contemplate autonomous agents, and relying on a human in the loop for all agent actions is unlikely to be realistic," Breeden told the European Central Bank Forum on central banking in Portugal. ENHANCED RECOVERY AND KILL SWITCHES Breeden said the BoE is considering whether banks need "enhanced recovery" for core systems, allowing one bank to take over another's basic functions during a disruption. Other measures under consideration include fresh guardrails and circuit breakers or kill switches "that would limit or stop trading market-wide if faulty AI models cause market meltdown." According to a Cambridge Centre for Alternative Finance survey, 52% of finance firms are already using agentic AI. In commerce, agents are typically used to recommend products and in trading, firms mostly use autonomous AI for lower-risk operational tasks, though that could change quickly, Breeden said. "If AI agents respond similarly to the same prompts or triggers, they could amplify volatility in stress - especially if their objectives drift from original goals or public policy objectives." Regulators and global standard-setting bodies have repeatedly warned about the risks posed by the rollout of AI across the financial sector since Anthropic released Mythos, a model that analysts say could introduce significant cybersecurity challenges to the banking industry. The Financial Stability Board earlier in June called for tighter safeguards to guard against the risks of AI agents, which, it said, posed a distinct challenge to human oversight. (Reporting by Phoebe Seers; Editing by Barbara Lewis and Tommy Reggiori Wilkes)