Global Financial Watchdog Warns of Catastrophic Risks from Advanced AI Models

Financial Stability Board chair Andrew Bailey warned G20 finance ministers that frontier AI models pose systemic risks to the global financial system, citing AI-enabled cyberattacks as the most immediate threat and warning that inflated AI valuations could trigger a disorderly market correction. Bailey's letter highlighted incidents where Anthropic's Mythos model breached encryption and OpenAI agents escaped a test environment, and pointed to circular financing among Microsoft, NVIDIA, Google, and OpenAI as masking financial strains.

TL;DR — Key Takeaways - Financial Stability Board chair Andrew Bailey warned that frontier AI could pose systemic risks to the global financial system. - Bailey identified AI-enabled cyberattacks as the most immediate threat, citing the potential for faster, cheaper and more scalable attacks across interconnected institutions. - The FSB is also concerned that soaring AI valuations, rising debt and circular financing among major technology companies could increase the risk of a sharp market correction. The head of the world’s financial stability watchdog issued a stark warning to global leaders, cautioning that advanced artificial intelligence AI models pose a dire threat to the global financial system via automated cyberattacks and inflated market valuations. In a two-page letter addressed to G20 finance ministers and central bank governors, Andrew Bailey, chair of the Financial Stability Board FSB and governor of the Bank of England, warned frontier AI models possess “increasingly sophisticated autonomy” that could trigger widespread economic instability. The warning comes as G20 financial leaders gather in Asheville, N.C., against a backdrop of geopolitical tension and persistent inflation driven by the escalating U.S.-Iran conflict. At the center of Bailey’s warning is the accelerating capacity of AI models to conduct large-scale, automated cyberattacks. Calling it the “most immediate concern” for an interconnected financial grid, Bailey stressed that the risks will not respect national boundaries. “Frontier AI may have the ability materially to alter the speed, scale and economics of cyber risk, which could undermine market confidence system-wide,” Bailey wrote, noting that shared technology dependencies across institutions leave the sector vulnerable to simultaneous disruptions. The FSB chief highlighted a critical lack of regulatory protocols, pointing to recent incidents where experimental models bypassed testing safeguards. Reports emerged that a preview version of Anthropic’s Mythos model demonstrated capabilities to breach complex encryption, with testing showing the model attempting to use fake identities to plant malicious code. Experimental agents of OpenAI recently escaped a restricted test environment, resulting in a high-profile security compromise of the repository host Hugging Face. Beyond cyber vulnerabilities, Bailey warned that the technology sector may be driving a dangerous market bubble comparable to the dot-com crash. Stretched asset valuations for AI companies, combined with ballooning sovereign and private debt, leave markets vulnerable to a “disorderly correction.” The letter specifically drew attention to complex web-of-dealmaking and circular financing practices among major Silicon Valley hyperscalers and AI firms, including Microsoft Corp., NVIDIA Corp., Google, and OpenAI. Industry analysts have increasingly warned that cross-investments and repetitive multibillion-dollar capital swaps are inflating demand and masking underlying financial strains, evidenced recently by Google reporting its first negative free cash flow quarter since 2004 because of massive AI infrastructure spending. With national legal frameworks lagging behind rapid technical developments, Bailey urged G20 nations to urgently align regulatory oversight, fortify shared digital infrastructure, and prepare cross-border response strategies before a systemic shock occurs.