The Bank of England’s deputy governor for financial stability just painted a picture of financial markets that should make anyone with a portfolio pay attention. Sarah Breeden, speaking at the European Central Bank Forum on Central Banking on June 30, warned that autonomous AI agents, the kind that can make trading decisions without human oversight, could accelerate cyber threats and drive correlated trading behaviors that amplify market shocks.
What Breeden actually said
Breeden’s address centered on what the industry calls “agentic AI,” meaning artificial intelligence systems that operate autonomously, executing decisions in financial markets without a human pressing the button. The concern isn’t theoretical. It’s about what happens when these systems, trained on similar data and optimized for similar objectives, start behaving in lockstep.
Breeden also flagged the acceleration of cyber threats as a parallel concern. Autonomous AI doesn’t just trade faster. It also creates new attack surfaces and can be weaponized in ways that legacy cybersecurity frameworks weren’t designed to handle.
The Bank of England is actively building computational simulations to model how AI agents behave in financial scenarios. The BoE is also looking at the design of “objective functions,” the goals that AI agents are programmed to optimize for.
