Bank of England reviews rules for agentic AI in finance
Agentic AI in financial services faces Bank of England scrutiny as rules lag See how regulators may curb cyber, trading, and operational risks in finance
The Bank of England is reviewing whether current regulations are adequate for the use of agentic AI in financial services, including payments, trading, cybersecurity, and operations. Deputy Governor Sarah Breeden said existing frameworks were not designed for AI systems that can act with limited human input.
Speaking at the European Central Bank Forum in Portugal, Breeden said it is unlikely that every action by these systems can be supervised directly by people. She said the technology is already being used in some finance tasks, with most adoption so far concentrated in internal processes and lowerrisk operational work.
The Bank of England is especially focused on cyber resilience and market stability. Breeden warned that AI tools could help defend firms, but could also make attacks more effective if used by malicious actors. Regulators are also considering safeguards such as circuit breakers and kill switches to limit disruption if AIdriven trading behaves unexpectedly. The review comes as global regulators, including the Financial Stability Board, also examine how to manage risks from autonomous AI systems.