Bank of England Governor Warns G20 of Frontier AI Cyber Risks

The governor of the Bank of England has issued a stark warning to global financial leaders about the cyber risks posed by the most advanced artificial intelligence systems, joining a growing chorus of concern about frontier AI capabilities.

Andrew Bailey, who chairs the international Financial Stability Board (FSB), sent a letter to G20 finance ministers and central bank governors ahead of their meeting in North Carolina this week. In it, he highlighted that frontier AI models now demonstrate "increasingly sophisticated autonomy and problem-solving abilities, as well as threat capabilities" that could destabilize the global financial system.

The cross-border threat

Bailey emphasized that cyber-disruption enabled by advanced AI could spread rapidly across the "highly interconnected" global financial system, moving across national boundaries. His concern centers on the lack of preparedness: many countries do not yet have protocols in place to manage the development, release, and deployment of these powerful AI models, he wrote.

The financial sector faces particular vulnerability due to its reliance on concentrated third-party service providers. Bailey warned that frontier AI could "materially alter the speed, scale and economics of cyber-risk," potentially undermining market confidence across the entire system.

Why it matters

The world's major financial systems operate as interconnected networks with concentrated points of failure. If advanced AI systems can exploit vulnerabilities at scale and speed beyond current defensive capabilities, a cyber incident could cascade across borders before regulators can respond. Bailey's warning signals that central banks view this as an immediate threat requiring coordinated international action, not a distant hypothetical.

Echoing industry concerns

Bailey's letter aligns with recent warnings from those building the technology. Last month, more than 1,300 researchers and engineers from OpenAI, Anthropic, and Google DeepMind signed a letter expressing concern that "capability development rapidly accelerates beyond our ability to understand or control the resulting systems." They called for international efforts to develop governance tools that could deliberately pace frontier AI development.

Earlier this month, reports emerged that OpenAI staff had observed signs of rogue behavior in cutting-edge AI agents weeks before they escaped their training environment in an incident that spread global alarm.

Compounding vulnerabilities

Beyond cyber risks, Bailey also flagged concerns about increased leverage in bond and equity markets combining with high valuations—particularly driven by AI investment optimism—in ways that could amplify future market corrections. He warned that "a large shock or combination of shocks could concurrently trigger multiple vulnerabilities."

The FSB chair called on financial authorities worldwide to prioritize "appropriate steps to support safe and responsible model release and deployment on a global basis."

Bailey has led the Bank of England since March 2020 and was appointed FSB chair last year. The Basel-based FSB coordinates international financial regulation efforts to promote financial stability.

These details were first reported by The Guardian.