Bailey Warns Against Weakening AI Regulation in Finance

Bank of England governor Andrew Bailey has been warning against weakening AI regulation in finance, arguing that the technology's rapid advancement demands stronger safeguards. In a series of public interventions over the summer, Bailey emphasized that frontier AI models are getting powerful enough to pose genuine cyber threats to the financial system.

The current testing infrastructure is not keeping pace with these advancements, and Bailey has called for more rigorous pre-deployment and post-deployment testing protocols before these models are trusted with systemic responsibility. He has also highlighted the vulnerability of UK banks in accessing Anthropic's Mythos model, which he identified as a significant weakness in managing cyber threats.

Bailey's push for international coordination is necessary to address this issue, as no single regulator can compel access to frontier models built by foreign companies. This requires diplomatic-level agreements, such as those hammered out at G20 summits. Bailey has also pointed out that existing vulnerabilities in the financial system, including high leverage and concentrated asset valuations, could be amplified by advanced AI.

The governor's implicit argument is that the cost of compliance with stricter regulations is lower than the cost of a major AI-enabled cyber incident, which could trigger a cascading confidence crisis. The Bank of England already mandates that banks demonstrate cyber resilience through stress tests and penetration testing.