China banks buffer against AI contagions as US sweats over Anthropic’s Mythos

With US Treasury Secretary Scott Bessent and Federal Reserve chair Jerome Powell reported to have held urgent talks over Anthropic’s latest artificial intelligence model, experts are warning that immediate cybersecurity assessments are needed to avert potential economic damage running into the hundreds of billions of US dollars.

Across the Pacific, however, analysts said China’s banks were taking a more cautious approach, as Beijing prioritises financial stability and intensifies monitoring of any potential contagion from the US market.

“It is urgent that cybersecurity researchers everywhere evaluate what a model of this capability would do to the economy and critical infrastructure,” said Simon Goldstein, a philosophy professor at the University of Hong Kong who specialises in AI safety.

Bessent and Powell held a short-notice meeting on Tuesday with the heads of major US banks to address concerns surrounding Anthropic’s model, Mythos, ensuring that the banks were aware of the risks posed by Mythos and potential future models, and were taking precautions to defend their systems, according to a Bloomberg report on Thursday night, US time.

Emergency summits between the Treasury and the Fed are relatively uncommon. While the two agencies are in regular contact, hastily convened high-level discussions are more often seen during moments of acute systemic stress, such as the 2008 global financial crisis and the 2020 pandemic-induced market crash.

Goldstein added that there was already speculation that Mythos could cause hundreds of billions in damages if it were released, citing an estimate by Ryan Greenblatt, chief scientist at Redwood Research, in a social media post.