Major tech firms quietly collaborate on a new initiative
Federal Reserve chairman Jerome Powell and Treasury secretary Scott Bessent met with the heads of major US banks this week to discuss potential cyber risks posed by Anthropic’s newly released artificial intelligence model, Claude Mythos Preview, CNBC confirmed Friday.
The meeting was held Tuesday at the Treasury Department, where bank executives were already gathered in Washington, D.C., for a Financial Services Forum board meeting, according to people familiar with the matter who were not authorized to speak publicly about the confidential discussions.
Bank of America CEO Brian Moynihan, Citigroup CEO Jane Fraser, Goldman Sachs CEO David Solomon, Morgan Stanley CEO Ted Pick, and Wells Fargo CEO Charlie Scharf were in attendance, Bloomberg reported. JPMorgan Chase CEO Jamie Dimon was the only major banking chief unable to attend.
Earlier in the week, Anthropic rolled out Claude Mythos Preview in a limited release, citing concerns that hackers could exploit its capabilities. JPMorgan Chase was among the initial launch partners for the cybersecurity initiative, known as Project Glasswing, alongside Apple, Google, Microsoft, and Nvidia.
Ahead of the model’s release, Anthropic briefed senior US government officials on its “offensive and defensive cyber applications.” An Anthropic official also told CNBC that the company had been in “ongoing discussions” with the Cybersecurity and Infrastructure Security Agency and the Center for AI Standards and Innovation.
“The dangers of getting this wrong are obvious, but if we get it right, there is a real opportunity to create a fundamentally more secure internet and world than we had before the advent of AI-powered cyber capabilities,” CEO Dario Amodei wrote on X.
The concerns surrounding Mythos stretch back to November, when Anthropic disclosed what it described as “the first reported AI-orchestrated cyber espionage campaign,” detecting suspicious activity in mid-September 2025 that investigators linked with high confidence to a Chinese state-sponsored group that used Claude’s agentic capabilities to execute cyberattacks themselves.
Public alarm escalated further in late March, when details about Mythos leaked through an unpublished blog post first reported by Fortune, which Anthropic attributed to human error within its content management system. Cyber stocks slumped in the days that followed.
