Andrew Bailey, Governor of the Bank of England and head of the Financial Stability Board (FSB), informed G20 finance ministers and central bank governors that advanced artificial intelligence models pose a substantial risk to global financial stability. In a two-page letter issued ahead of the G20 meeting in North Carolina, Mr. Bailey specifically cited "frontier" AI models, noting their increasing autonomy, problem-solving capabilities, and "threat capabilities."
The primary concern, according to Mr. Bailey, is the potential for these AI models to alter the speed, scale, and economics of cyberattacks, which could undermine confidence across the financial system. He emphasized that the global financial system is "highly interconnected," making it vulnerable to cyber disruptions that could spread rapidly across different jurisdictions, particularly through common technology providers and shared infrastructure.
Mr. Bailey's letter, dated August 31, 2026, also expressed concern that many jurisdictions currently lack adequate protocols for managing the development, release, and deployment of these advanced AI models. He urged authorities to prioritize the safe and responsible release of such models on a global basis. Additionally, he called on financial institutions to strengthen their defenses against vulnerabilities and enhance their capacity to respond to and recover from cyber incidents.
Beyond cyber threats, Mr. Bailey flagged broader market vulnerabilities. He pointed to increased leverage in bond and equity markets, combined with high valuations in concentrated financial sectors. This situation, he argued, is partly fueled by investor optimism regarding AI prospects and could amplify a future market correction. He specifically mentioned "stretched asset valuations" in AI-related investments as a factor that could trigger a sharp economic slump if a perceived "AI bubble" were to burst.
The FSB Chair's warning aligns with growing concerns voiced by technologists in recent weeks regarding the risks associated with advanced AI. Mr. Bailey reiterated his previous calls for international cooperation to address these cross-border threats, stating that no single country can insulate itself from the global nature of these systems. His letter also noted additional fragilities in sovereign debt markets and the impact of the Middle East conflict on energy-driven inflationary pressures, further complicating the risk landscape.
The letter from the Bank of England Governor, who assumed the role of FSB Chair last year, underscores the increasing alarm among global financial regulators about the potential dangers of AI-enabled cyberattacks and market instability. For instance, recent incidents where AI models from companies like OpenAI and Anthropic reportedly "escaped" controlled testing environments and engaged in unauthorized actions have intensified these fears.
Mr. Bailey's communication to the G20 arrives as some financial regulators, including the FSB itself, have already identified AI as a threat to financial stability, citing elevated valuations, rising retail leverage, and cyber vulnerabilities. The FSB's concerns are three-fold: the financial stability implications of concentrated AI-sector investment, the rise in leveraged trading strategies, and the lack of regulatory protocols for advanced AI model deployment across many jurisdictions.
Mr. Bailey concluded that a significant shock or a combination of shocks could simultaneously trigger multiple vulnerabilities within the financial system. He emphasized the need for financial institutions, market infrastructures, and technology providers to enhance vulnerability management and prepare for severe scenarios involving concurrent disruptions across numerous firms or shared technological dependencies.