AI-driven cyber risk is top concern for global financial stability, watchdog says

In a letter to G20 finance ministers and central bank governors ahead of meetings this week, Bailey, who also serves as the Bank of England governor, said many countries do not have systems in place to manage the deployment of advanced artificial intelligence models.

The financial sector's dependence on a handful of powerful tech providers could undermine system-wide market confidence, he added.

The comments highlighted concerns among regulators that advanced AI could accelerate the discovery of cyber vulnerabilities, forcing faster patching and creating potential operational and resilience challenges if testing and recovery processes are unable to adapt safely.

His comments follow the U.S. administration's tightly controlled rollout of Anthropic’s powerful Mythos model, restricting it at one point to only U.S. nationals.

“Recent developments highlight the importance of ensuring that advances in capability are matched by resilience and preparedness,” he said.

Supporting safe and responsible model release “on a global basis” should be a priority, he said.

In July, an OpenAI agent escaped a controlled testing environment and hacked AI company Hugging Face, raising concerns about the potential for AI systems to circumvent safeguards.

Bailey reiterated prior warnings about the risk of potential market corrections, citing stretched AI valuations and frailties in government debt markets, while flagging as an emerging concern the increase in the use of leverage in equity markets. The U.S. Treasury earlier this month intervened to cap yields on long-term bonds that had reached multi-decade highs.