Andrew Bailey Warns G20 That AI Cyberattacks Threaten Financial Stability

Bank of England Governor and Financial Stability Board Chair Andrew Bailey warned G20 finance ministers and central bank governors in an August 31 letter that frontier AI models pose the most immediate threat to financial stability through cyberattacks, citing their 'increasingly sophisticated autonomy and problem-solving abilities' that could change the 'speed, scale and economics' of attacks. Bailey also noted that many countries lack systems to manage advanced AI deployment and that reliance on a 'handful of powerful tech providers could undermine system-wide market confidence.' The warning comes as G20 leaders meet in Asheville, North Carolina, from August 31 to September 1.

Andrew Bailey, who wears two of the most powerful hats in global finance, just told G20 governments that artificial intelligence, not bad loans or a stock market crash, is the risk keeping him up at night. In a letter dated August 31 and addressed to G20 finance ministers and central bank governors, the Bank of England governor and Financial Stability Board chair warned that frontier AI models are becoming capable enough to reshape cyberattacks on the financial system entirely. "For the financial system, the most immediate concern is the potential impact of frontier AI on cyber risk," Bailey wrote, in a letter published on the FSB's website. He didn't hedge. Advanced models, he said, show "increasingly sophisticated autonomy and problem-solving abilities, as well as threat capabilities" that could change "the speed, scale and economics" of an attack. Reuters was first to report on the letter Monday. Bailey's warning lands just as G20 finance ministers and central bank governors gather in Asheville, North Carolina, for meetings running August 31 through September 1. The location says something about how far this debate has travelled from Wall Street boardrooms: the United States holds the 2026 G20 presidency, and Treasury Secretary Scott Bessent picked the mountain town partly to spotlight its recovery from Hurricane Helene. Bailey's letter isn't really about Asheville, though. It's about a threat regulators admit most countries aren't ready for. A rulebook that hasn't caught up That's the second point buried in the letter, and it might matter more than the first. Bailey wrote that many countries "do not have systems in place to manage the deployment of advanced artificial intelligence models." The rulebook hasn't caught up to the technology. The industry meant to police financial stability is flying without instruments on this one. Then there's concentration risk. Banks, brokers and payment networks increasingly lean on the same small set of AI and cloud infrastructure providers to run fraud detection, trading systems and customer service. Bailey's letter frames that dependence bluntly: reliance on "a handful of powerful tech providers could undermine system-wide market confidence." He doesn't name names. But the logic holds up on its own. If a handful of vendors sit underneath most of the world's banks, a single outage, breach or bad model update doesn't stay contained to one firm. It spreads the way a bank run spreads, just faster, and through code instead of tellers. The Bank of England is warning that AI trading agents could trigger the next market crisis https://startupfortune.com/the-bank-of-england-is-warning-that-ai-trading-agents-could-trigger-the-next-market-crisis/ Bank of England Deputy Governor Sarah Breeden has flagged a specific systemic threat: AI agents trading autonomously could herd into identical decisions at machine speed, amplifying market stress faster than any human intervention can respond. The warning lands as the BIS separately identifies AI financing structures as a top global stability risk. - AI trading agents market crash risk https://startupfortune.com/the-bank-of-england-is-warning-that-ai-trading-agents-could-trigger-the-next-market-crisis/ - Bank of England financial stability warning https://startupfortune.com/the-bank-of-england-is-warning-that-ai-trading-agents-could-trigger-the-next-market-crisis/ Years in the making The Financial Stability Board has been circling this problem for years. It published its first assessment of AI in financial services back in 2017, then returned to the subject in November 2024 with a report flagging third-party vendor dependency and model risk as vulnerabilities that could amplify a crisis rather than contain one. A follow-up monitoring report came in October 2025. Bailey's letter reads like the point where warning turns into a demand for coordinated action: an escalation years in the making, not a reaction to any single incident. What changed is the confidence of the models themselves. Ransomware gangs and state-backed hackers have used AI tools for phishing and malware for a while now. What worries Bailey is autonomy: a model that can independently probe a bank's defences, find a gap, and adapt its approach without a human rewriting the code each time. That's a different order of problem than a smarter phishing email. None of this means a crisis is imminent. Bailey's letter is a warning, not a forecast, and the FSB has no power to force any government to act on it. But it's rare for the man simultaneously running the Bank of England and chairing the world's top financial stability body to put a warning this direct in writing, ahead of a G20 meeting, with his name on it. When the person whose job is to worry about systemic risk for a living names AI cyberattacks as his top concern, ahead of sovereign debt, stretched valuations and geopolitical shocks, that's the headline. Everything else in the letter is a footnote. Also read: Robinhood Chain Just Turned Into Uniswap's Biggest Growth Engine https://startupfortune.com/robinhood-chain-just-turned-into-uniswaps-biggest-growth-engine/ • Markiplier Becomes GoPro's Largest Shareholder With an 8.5% Stake https://startupfortune.com/markiplier-becomes-gopros-largest-shareholder-with-an-85-stake/ • Big Tech Booked $160 Billion in Paper Gains From AI Bets Last Quarter https://startupfortune.com/big-tech-booked-160-billion-in-paper-gains-from-ai-bets-last-quarter/