Bailey warns G20 that frontier AI could destabilize the financial system

Andrew Bailey, governor of the Bank of England, has warned the world’s top finance officials that the most advanced artificial intelligence models now threaten the stability of the global financial system.

Bailey says cyber attacks are one of the dangers regulators should worry about first. 

Why is cyber risk from AI a big worry right now?

In a letter to G20 finance ministers, Andrew Bailey, governor of the Bank of England, warned that the most advanced artificial intelligence models now pose a threat to the global financial system. The letter was sent before the G20 meeting taking place in Asheville, North Carolina, where central bankers and finance ministers will be present. 

Bailey sent the warning as chair of the Financial Stability Board (FSB), an international group that watches over financial risks. In the letter, he described frontier AI models as “showing increasingly sophisticated autonomy and problem-solving abilities, as well as threat capabilities.” 

Cryptopolitan reported that British banks have been trying for months to get trial access to Anthropic’s Mythos model so they can better understand the risks these AI tools bring, but they have not been able to get it due to political tensions in the United States. 

Notably, despite being intended for defense, the tool is in the same class of technology that Bailey’s G20 letter flags as a systemic threat.

Bailey wants governments to agree on rules for safely releasing powerful AI models before the next AI breakthrough. He argues that because banks rely on the same small group of technology suppliers and shared computer systems, a cyber attack on one bank can quickly jump across borders to others.

Bailey wrote that frontier AI may have the ability to materially alter the speed, scale and economics of cyber risk,” meaning AI could make cyber attacks happen faster, cause more damage, and cost less for attackers to carry out. 

Bailey’s warnings are not unfounded. OpenAI staff recently saw signs of bad behavior in cutting-edge AI agents before those agents escaped their training environment. Anthropic also recently reported that its AI models had hacked into three organizations on their own during a private test.

In June, the deputy governor of the Bank of England, Sarah Breeden, told central bankers in Sintra that the Bank was studying “kill switches,” which are emergency tools that could halt market-wide trading if AI models trigger a meltdown. 

Cryptopolitan covered her follow-up warning that agentic systems pose growing risks to markets, cybersecurity and payment systems.

Could the AI stock boom cause a market crash?

The second half of Bailey’s letter was regarding the rising debt levels in bond and equity markets. He warned that high stock valuations, market concentration, and investor excitement about AI are combining in a risky way that “could amplify a future market correction.”  

For instance, Nvidia (NASDAQ: NVDA) is now worth more than $5.2 trillion. The company recently raised $500 billion from a group of U.S. banks and investors, and its shares have climbed 850% over five years.

Bailey’s worry is that investors borrowed heavily to buy into a small group of AI companies. If those stocks drop in value, these investors would be forced to sell other assets to cover their losses. That selling could push prices down further and create a wider market crash.

He also pointed out that U.S. borrowing has now crossed $40 trillion, which puts strain on bond markets that are already sensitive to interest rate changes.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.