The Bank of England is reviewing whether its existing rules can cover the use of agentic AI in finance. This includes areas like payments, trading, cybersecurity, and operations.
Deputy Governor Sarah Breeden said the current regulatory frameworks were not designed for AI agents that can act without direct human instruction. She made these remarks at the European Central Bank Forum on central banking in Portugal.
Why existing rules may not work for agentic AI
According to Artificial Intelligence News, Breeden stated that existing frameworks were not built to contemplate autonomous agents in payments, trading, and operational functions. She added that relying on human oversight for every action by these systems is unlikely to be practical.
The Bank of England's review comes as agentic AI systems are already being used in financial workflows. These systems can make decisions and carry out tasks independently, without waiting for human approval at every step.
What agentic AI means for finance
Agentic AI refers to systems that can make decisions and carry out tasks independently. In finance, such systems are already being used in areas like automated trading, payment processing, and cybersecurity threat detection.
As reported by Reuters, Breeden signaled that new rules may be needed to govern agentic AI. The Bank of England is now examining whether existing frameworks can adequately cover these autonomous systems or if regulatory reform is required.
Our Take: A necessary review for a fast-changing technology
The Bank of England is right to ask this question now. Agentic AI is not a future possibility — it is already being used in finance. The fact that existing rules were not designed for systems that act without human instruction is a real gap that needs to be addressed.
In our view, the key challenge is finding the right balance. Too much regulation could slow down innovation and put UK finance at a disadvantage. Too little could leave the system exposed to risks that regulators do not fully understand yet.
Breeden's point about human oversight being impractical is important. If every AI action needs a human to check it, the whole point of using these systems is lost. But letting them run completely unchecked is also not acceptable. The Bank of England will need to find a middle ground — and it is good that they are starting this conversation now.
