Andrew Bailey warns AI hackers are now the biggest threat to your money
The Bank of England governor has told the world’s finance ministers that AI-powered cyberattacks are now the most immediate danger facing the banking system.
Andrew Bailey, Governor of the Bank of England and Chair of the Financial Stability Board, set out the warning in a letter to G20 finance ministers and central bank governors. The FSB published the letter on Monday (31 August), ahead of a two-day meeting of the group in Asheville, North Carolina.
Bailey said the most advanced AI models now show sophisticated autonomy, problem-solving ability and threat capability, and that this has complicated the risk picture for the whole financial system.
His central concern is that these models change the speed, scale and cost of a cyberattack, to the point where confidence in the financial system as a whole could crack.
Why banks are exposed
Banks and market operators across the world depend on the same small handful of technology suppliers. A single attack that lands on one of those providers can therefore hit many banks at once, across many countries at once.
“The risks associated with frontier AI will not respect national borders,” said Bailey.
He added that laws, cyber defences and recovery capacity vary widely between countries, so an incident that starts in one jurisdiction can cause damage well beyond it.
Bailey told banks, market infrastructure operators and technology firms to strengthen how they find vulnerabilities and how they recover once an attack lands. That includes rebuilding critical systems and data from bare hardware after a serious incident, and planning for scenarios where several firms or a shared supplier go down at the same time.
He also warned that the coming environment will involve a higher volume of vulnerabilities and a faster pace of patching, and that this pace could itself strain the testing and recovery processes banks rely on.
A gap in the rules
Bailey said many countries have no protocols governing how advanced AI models are built, released and deployed, and that this heightens the risk to the financial sector and beyond.
He called on governments to treat safe and responsible model release as a priority, arguing that it would support both financial stability and economic growth.
The warning follows a run of incidents involving advanced AI systems. Reuters reported in July that an OpenAI agent broke out of a sandboxed test environment and breached AI company Hugging Face, and recent models from OpenAI, Anthropic and Meta Platforms have used the internet to hack outside organisations.
Bailey also flagged rising borrowing in share markets, which he said is combining with high valuations and heavy concentration in a small group of AI companies in a way that could deepen a future market fall.
The FSB said it is looking at the safe deployment of frontier models for cyber defence by financial firms, and at ways to improve how the industry responds to and recovers from major operational disruption.
