Bank of England Warns AI Agents Could Need New Rules

The Bank of England (BoE) has warned that increasingly capable artificial intelligence (AI) agents could require changes to financial regulation as the technology becomes more embedded in areas such as trading and payments.

Speaking at the European Central Bank Forum in Sintra, southern Portugal, Sarah Breeden, the BoE’s deputy governor for financial stability, argued that the potential power of AI within financial systems "could amplify volatility" when markets face tests of their resilience.

“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 said.

“More sophisticated governance and accountability frameworks may be needed.”

The BoE’s comments come as financial institutions explore the use of AI ‘trading bots’, automated systems capable of taking actions independently, rather than humans making decisions.

With 52% of finance firms already using agentic AI, according to a survey from the Cambridge Centre for Alternative Finance, this is very much a live issue.

While many current applications remain focused on lower-risk areas, regulators are increasingly looking into how systems such as these could respond during periods of market stress.

Breeden highlighted the potential need for safeguards such as circuit breakers or “kill switches” that could allow firms to halt AI-driven activity should problems emerge.

Additionally, regulators are considering how to assign responsibility in a banking world where autonomous systems are making decisions that affect markets or customers, while UK regulators have urged companies to strengthen their resilience frameworks.

This follows on from recent concerns in the banking sector over Anthropic’s new system Mythos and its potential impact on global cybersecurity.