US Treasury Sounds Alarm on AI Cyber Risks with Emergency Bank Meeting

The US Treasury secretary, Scott Bessent, recently convened a meeting with major American bank chiefs in Washington to address concerns over the cyber risks posed by Anthropic's latest AI model, Claude Mythos. This gathering was attended by prominent figures, including Jerome Powell, chair of the Federal Reserve, and several bank CEOs, such as David Solomon of Goldman Sachs, Brian Moynihan of Bank of America, and Charlie Scharf of Wells Fargo.

The meeting was prompted by a recent leak of Claude's code, which led Anthropic to publish a blog post warning that AI models have surpassed human capabilities in finding and exploiting software vulnerabilities. The startup expressed concerns that the fallout from these vulnerabilities could have severe consequences for economies, public safety, and national security.

Cyber Risks and Financial Stability

The guest list for the meeting focused on heads of systemically important banks, whose operations are crucial to maintaining financial stability. The meeting was reportedly called while bank bosses were already in Washington for a lobby group gathering. JP Morgan's Jamie Dimon was invited but unable to attend, and in his annual letter to shareholders, he warned that cybersecurity remains one of the biggest risks and that AI will likely exacerbate this issue.

Anthropic's Mythos model has exposed thousands of vulnerabilities in software and popular applications, prompting the company to limit the release of the new model to a select group of businesses, including Amazon, Apple, and Microsoft. This is the first time Anthropic has restricted the release of a product, highlighting the severity of the cyber risks posed by the AI model.

The news comes amid concerns that hackers could use such tools to figure out passwords or crack encryption, compromising data safety. The company revealed that the oldest vulnerabilities uncovered by Mythos were up to 27 years old, and none of them had been noticed by their creators or tech monitors before being identified by the AI model.

The meeting took place weeks after the US government designated Anthropic as a supply chain risk, allegations that the company is disputing in court. The Federal Reserve, Anthropic, and the US banks declined to comment on the matter, while the Treasury did not respond to requests for comment.