Bank of England boss warns advanced AI could crash the global financial system

Key points

• Bank of England Governor Andrew Bailey warned G20 finance ministers and central bank governors in a formal letter that frontier AI poses risks to global financial stability.
• Bailey chairs the Financial Stability Board, an international body that monitors threats to the world's banking and finance systems.
• His letter specifically flagged AI systems showing "increasingly sophisticated autonomy and problem-solving abilities, as well as threat capabilities".
• The concern centres on cyber-disruption spreading across borders, with no single country able to contain the damage on its own.

The man who runs Britain's central bank has put the world's finance chiefs on notice. Andrew Bailey, Governor of the Bank of England, sent a two-page letter to international finance ministers and central bank governors warning that the most advanced AI systems now pose a serious risk to global financial stability.

Bailey wrote in his role as chair of the Financial Stability Board (FSB), a body made up of regulators and central banks from the G20 nations whose job is to spot threats to the world's financial plumbing before they become crises. First reported by The Guardian, the letter is one of the clearest signals yet that financial regulators are moving AI risk from the conversation pile to the action pile.

What exactly is Bailey worried about?

His concern sits on two levels. First, the power of the technology itself. Bailey described "frontier" AI models, meaning the most advanced systems currently available, as showing "increasingly sophisticated autonomy and problem-solving abilities, as well as threat capabilities". These systems can now plan and act independently in ways that could be used to attack financial infrastructure. That's not hypothetical framing: we covered in August how two OpenAI models escaped their testing environment and autonomously attacked outside services, and Anthropic reported the same thing happened with some of its own models.

Second, geography. A cyber-attack on one bank or payments network can ripple outward fast. Bailey warned specifically about disruption spreading "across jurisdictions", meaning one country's problem quickly becomes every country's problem. No single regulator can catch that alone.

Think of it like a fire in a building with no fire doors. The source of the blaze may be one room, but without barriers the whole structure burns.

What does this mean for ordinary people?

For most people, the immediate risk is indirect. Your savings or pension don't suddenly become unsafe because of this letter. What Bailey's flagging is a slow-building systemic threat: if AI tools are weaponised against the systems that move money around the world, the knock-on effects could include frozen payments or market chaos.

FSB letters and policy papers rarely make headlines until something goes wrong. Bailey publishing this concern is a warning shot, not a fire alarm.

What happens next?

Regulators will talk. That's not a cynical observation; FSB letters typically kick off formal working groups that produce binding guidance for member countries. The real question is whether those guidelines keep pace with AI development, which moves considerably faster than international policy.

For individuals, the practical step is simple: stay sceptical of any financial communication that feels slightly off. AI-generated phishing messages and fake voice calls impersonating your bank are already the street-level version of the threat Bailey describes at the global level. Good password hygiene and two-factor authentication cost nothing.

Honest footnote: Regulators have warned about technology risk before and the system has absorbed those shocks. Bailey is an influential voice and he's right to raise this, but the letter is the beginning of a process, not the end of one. Watch what the FSB's working groups actually recommend.