Sarah Breeden, the Bank of England’s deputy governor for financial stability, warned that autonomous AI systems pose a growing threat to financial markets, cybersecurity, and payment infrastructure, urging central banks worldwide to look into better regulations. Speaking at the European Central Bank’s annual forum in Sintra, Portugal, she explained that agentic AI is advancing faster than regulators anticipated, noting that in 2019 the length of software tasks leading AI models could complete was doubling every seven months, but by 2024, the doubling was happening every four months. Breeden emphasized three phases of AI development, stating that now agentic systems can plan and execute sequences of decisions with zero human oversight, which could lead to AI agents trading securities, processing payments, and responding to cyber threats. She singled out cybersecurity as the most critical issue, citing the UK government’s AI Security Institute’s finding of a huge improvement in agentic AI capabilities, warning that malicious use “materially increases the chance of attacks that could harm financial stability.” Breeden suggested autonomous trading tools may need built-in “kill switches.” Additionally, the Bank’s Financial Policy Committee concluded in April that debt financing for AI infrastructure was increasing rapidly, meaning a sudden drop in AI-related asset valuations could cause a massive ripple through credit markets.