The Bank of England on Tuesday signalled the need for bespoke rules to govern agentic artificial intelligence, warning that autonomous agents could expose gaps in financial regulation and potentially trigger market-wide disruption. Deputy Governor Sarah Breeden said existing frameworks were not built for systems that make decisions without human oversight, and the central bank is now weighing market-wide kill switches to contain breakdowns.
"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.
Gaps in existing frameworks
Breeden said the BoE is considering whether banks need "enhanced recovery" capabilities that let one bank take over another's basic functions during an outage. Other measures include fresh guardrails and circuit breakers that "would limit or stop trading market-wide if faulty AI models cause market meltdown."
Adoption surging, risk scenarios emerging
According to a Cambridge Centre for Alternative Finance survey, 52% of finance firms are already using agentic AI. In commerce, agents typically recommend products, while trading desks use them mostly for low-risk operational tasks. That could change fast, Breeden noted, as systems grow more autonomous.
"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," she said.
Global regulators sound alarm
The Financial Stability Board earlier in June called for tighter safeguards against risks from AI agents, which it said posed a distinct challenge to human oversight. The warnings have intensified since Anthropic released Mythos, a model that analysts believe could introduce significant cybersecurity challenges to banking.
Why this matters for finance professionals
The BoE's shift signals that financial institutions will soon face more prescriptive expectations around AI governance. Compliance, risk, and technology teams will need to prepare for audit and oversight of agentic systems, including how kill switches and recovery plans work in practice. As agentic adoption grows, expertise in model risk management and operational resilience will become central to maintaining regulatory standing.
Your membership also unlocks:
