Bank of England reviews rules for agentic AI risks

The Bank of England is reviewing its regulatory frameworks to address risks from agentic AI systems, Deputy Governor Sarah Breeden told the European Central Bank Forum. The central bank is considering enhanced recovery measures, including allowing one bank to take over another's functions during disruptions and implementing market-wide kill switches for faulty AI models. The review targets risks in payments, trading, and cybersecurity.

Editorial analysis: For AI/DS practitioners, central-bank scrutiny of agentic systems raises the bar on production safety, observability, and cross-firm recovery planning. According to Reuters, Deputy Governor Sarah Breeden told the European Central Bank Forum that "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." Reuters reports the Bank of England is considering "enhanced recovery" measures that could allow one bank to take over another's basic functions during disruption and market-wide circuit breakers or "kill switches" if faulty AI models cause market meltdown. The original RSS description highlights affected areas including payments, trading and cybersecurity. The Financial Times and other outlets similarly emphasised the "kill switch" language in coverage.