The Bank of England has warned that existing regulatory frameworks may not be able to oversee increasingly autonomous artificial AI systems.

Speaking at the European Central Bank Forum on Central Banking in Sintra, Portugal, Sarah Breeden said advances in AI were happening at such a pace that policymakers should expect "more frequent technology surprises" and prepare for the financial stability risks they could bring.

"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," the Bank of England's deputy governor said.

Breeden said the emergence of agentic AI could transform areas including payments, trading and financial operations, but also introduce new systemic risks.

She warned that AI-powered trading could increase market volatility if autonomous systems begin responding in similar ways during periods of uncertainty.

"If AI agents respond similarly to the same prompts or triggers, they could amplify volatility in stress," she said. “The financial stability question therefore broadens from whether firms can use models well to whether the system can also observe and contain their resulting behaviours.”

The deputy governor said the Bank is exploring whether additional safeguards may be required, including market-wide circuit breakers or "kill switches" capable of halting trading if faulty AI systems threaten financial stability.

Breeden also highlighted cyber security as one of the most immediate challenges facing financial institutions, warning that AI is rapidly increasing the capabilities of both attackers and defenders.

"It's on us to ensure that the next technology surprise does not become a test of financial stability," she said. “AI is uncovering many vulnerabilities to patch, including as global software providers and key UK financial institutions use the latest models.

“For financial stability, patching must happen quickly – across the financial sector, keythird-party technology providers, and wider national infrastructure such as energy and telecoms. That is no small task.”

The speech also examined the implications of AI-powered payments, with Breeden warning that regulators will need to address questions around customer consent, liability, and fraud as AI agents begin making purchases and payments on behalf of consumers.

"What these two examples – agentic commerce and agentic trading – both highlight is that, as AI capabilities increase, we must keep asking whether existing, technology-agnostic regulatory frameworks remain sufficient," she said.

Breeden added that central banks must also embrace AI themselves, using the technology to strengthen supervision, identify emerging risks and improve stress testing as autonomous systems become a larger part of the global financial system.

---