Bank of England governor urges global action on AI risks

Bank of England governor Andrew Bailey called for stronger international cooperation on frontier AI, warning that governments need shared safeguards to prevent powerful digital tools from reaching bad actors. Speaking before the annual Mansion House dinner in London, Bailey said countries must improve their understanding of how advanced AI models are introduced and strengthen coordinated safety testing before wider release.

Bailey said the US cannot meet its security goals alone because financial and digital systems are deeply interconnected. His remarks followed Donald Trump’s temporary ban on foreigners using some of Anthropic’s most powerful models, including Claude Mythos, a move that frustrated allies seeking a joint approach to AI governance. Bailey later told City executives that no country can seal itself off from cross-border systems.

The comments came alongside a Mansion House speech by chancellor Rachel Reeves, who defended her record before an expected change of Labour leadership. Reeves warned that renewed Middle East hostilities would continue to test UK economic resilience, while the yield on 10-year UK government debt rose above 5% before easing. Demis Hassabis also called for a US-led global AI watchdog to test advanced models and pause development where risks become too high.