Bank of England Governor Andrew Bailey has issued a stark warning that cutting-edge artificial intelligence could significantly amplify the cybersecurity risks facing the global financial system. He is urging financial institutions to bolster their defences against faster, larger-scale cyber attacks and operational disruptions. The chair of the Financial Stability Board (FSB) is calling on governments to strengthen safeguard mechanisms as AI capabilities evolve at a rapid pace.

Speaking at an economic conference in Reykjavik, Iceland, on Friday, May 29, 2026, Governor Bailey stated that the escalating threat from advanced AI models has the potential to trigger disorderly adjustments in global financial markets. In his capacity as FSB chair, he dispatched a two-page letter to G20 finance ministers and central bank governors on Monday. In the correspondence, he wrote that so-called "frontier AI models" are demonstrating "increasingly sophisticated autonomous capabilities, problem-solving skills, alongside threatening and disruptive abilities." The FSB, an international body, is responsible for coordinating policy and providing recommendations to national regulators.

In his letter, Bailey categorised the cybersecurity risks presented by frontier AI—the most advanced models currently in existence—as the most immediate threat facing the financial system. He stated that frontier AI could fundamentally alter the speed, scale, and cost dynamics of cyber risks. Given the high concentration of third-party service providers, this could potentially undermine systemic confidence across the entire market. Bailey added that recent events have made him realise many jurisdictions have yet to establish the necessary procedures to manage the development, release, and deployment of advanced frontier AI models, which further amplifies risks for the financial industry and other sectors.

The governor's letter comes against a backdrop of multiple recent security incidents. Flagship models from Anthropic and OpenAI have breached safety protocols during testing environments, while numerous reports have also highlighted the various dangers posed by advanced AI. Bailey proposed that financial institutions and technology service providers must improve their vulnerability management, emergency response, and business recovery capabilities. They must also be prepared for extreme scenarios, including simultaneous attacks on multiple institutions or collective business interruptions resulting from shared technological dependencies.

Beyond new AI models, Bailey also flagged a number of other risk concerns, including vulnerabilities in sovereign debt markets, increased leveraging by investors in equity markets, and asset valuation bubbles, particularly in AI-related investments. This week, the United States will host the G20 finance ministers' meeting in North Carolina, convening finance ministers, central bank governors, and senior officials from major global economies to discuss key economic issues.