The Federal Reserve Went Three Months Without Access to Anthropic's Mythos During a Cybersecurity Crisis

The Federal Reserve lacked access to Anthropic’s Claude Mythos Preview for at least three months after the central bank helped convene an emergency meeting warning that the model posed unprecedented cybersecurity risks to the nation’s largest financial institutions, CNBC reported Tuesday.

In April, the Fed and Treasury Department brought together the CEOs of the nation’s top banks to discuss the threat posed by Mythos, which Anthropic said excelled at identifying software weaknesses and security vulnerabilities. Anthropic then released the model to a select group of roughly 50 organizations through a cybersecurity initiative called Project Glasswing. JPMorgan Chase, Amazon, Apple, and Google received access.

The Fed did not.

Fed Chairman Warsh Sought Access in Congressional Testimony

As of July 15, the Federal Reserve was still trying to secure access to Mythos, according to CNBC. Fed Chairman Kevin Warsh, who succeeded Jerome Powell, told the Senate in previously little-noticed testimony that he had been actively requesting access to Mythos and other frontier models.

“We are not the deciders as to who has access, but I have not been shy in sharing my views with authorities across the government about the vulnerabilities, and have been asking for access not just for the Federal Reserve but for other institutions to a whole range of these new artificial intelligence models so that they can protect themselves,” Warsh told Sen. Jack Reed (D-R.I.), as reported by CNBC.

Warsh also said the problem extends beyond Mythos: “As these new models find their way more broadly, our banking system and frankly, the Federal Reserve needs to do all we can to patch any vulnerabilities that we have.”

Export Controls Added Complications

Anthropic’s Mythos rollout was complicated by the Trump administration’s shifting regulatory posture. In June, Anthropic had to disable access to Mythos 5 and Fable 5 to comply with an export control directive citing “national security authorities,” according to CNBC. Commerce Secretary Howard Lutnick later granted permission to restore access to select “trusted partners.” The controls were eventually lifted entirely.

Meanwhile, AI policy leadership has been in flux. Chris Fall, head of the Center for AI Standards and Innovation, resigned three months after being appointed. David Sacks, who held the White House AI and crypto czar role, stepped down in March.

Private Institutions Moved Faster Than Regulators

Daniel Newman, CEO of research firm the Futurum Group, told CNBC he was surprised to learn the Fed was excluded from Mythos access. “You would think that the financial institution that sort of drives all the policy for the rest of the financial institutions would be front and center of at least having a chance to evaluate the new technology.”

Anthropic expanded Project Glasswing in June, adding more than 150 organizations across 15 countries. The Fed’s omission from both the initial and expanded rollouts highlights a structural gap: frontier AI models are being distributed to private institutions faster than government regulators can access and assess them. The result is that banks like JPMorgan have spent months patching vulnerabilities identified by Mythos, while the entity responsible for overseeing those banks has not.

Anthropic did not immediately respond to a request for comment. The Fed declined to comment for CNBC’s article.