The Bank of England's second-in-command says AI trading agents may need new regulation before they cause a market feedback loop nobody can stop.
Deputy Governor Sarah Breeden raised the concern at the European Central Bank's annual forum in Sintra, Portugal. Her core worry is not a rogue algorithm going haywire — it is synchronization. If autonomous trading agents are trained on similar data and built around similar logic, they could all react the same way to the same signal at the same moment, amplifying a dip into a rout. Breeden indicated that existing regulatory frameworks may not be equipped to handle that kind of herd behavior when the herd moves at machine speed.
The warning lands at a moment when AI-driven trading tools are spreading fast across financial markets, and regulators are still writing the rules. A synchronized sell-off among human traders is bad; among thousands of agents executing in milliseconds, the feedback loop could outrun any circuit breaker designed for human-speed markets. That is the gap Breeden is flagging.
Central bankers have been warning about systemic AI risk in finance for a few years now, but the language is getting sharper — which usually means the concern is getting closer to becoming policy.