Fed Chair Jerome Powell, Treasury’s Scott Bessent Meet with Top Bank CEOs to Discuss Cybersecurity Risks from Anthropic’s AI Model ‘Mythos’
Washington, D.C., April 10, 2026 — Federal Reserve Chair Jerome Powell and Treasury Secretary Scott Bessent convened a closed-door meeting on Tuesday with leading chief executives from major banks to discuss emerging cybersecurity threats posed by the latest artificial intelligence model developed by Anthropic, known as Mythos. The meeting, held at the Treasury Department’s headquarters, reflects growing governmental concern over the potential risks advanced AI technologies pose to financial institutions and national security, CBS News has learned from sources familiar with the event.
Meeting Overview and Attendees
The session brought together top banking leaders alongside Powell and Bessent to address the implications of Anthropic’s Mythos, an AI model noted for its advanced generative capabilities. Although Jamie Dimon, CEO of JPMorgan Chase, was invited, sources confirmed he was unable to attend.
Earlier reports by Bloomberg News revealed that both Powell and Bessent sought the expertise of financial executives amid rising worries that cutting-edge AI systems could expose vulnerabilities across critical infrastructure sectors, including banking and finance.
Anthropic’s Mythos and Project Glasswing
Anthropic, the AI company behind the chatbot Claude, announced on Tuesday that it is initiating a collaborative project—named Project Glasswing—with several tech giants such as Amazon, Apple, and Nvidia. The initiative aims to harness the powerful capacities of Mythos for bolstering cybersecurity defenses.
The company has chosen not to make Mythos widely available to the public due to its ability to uncover significant weaknesses in major operating systems and web browsers. According to Anthropic, the rapid pace at which AI technology is developing may soon lead to widespread availability of comparable systems, including in the hands of actors who may not prioritize safe deployment.
In an official statement, Anthropic cautioned against the potential fallout from uncontrolled AI proliferation, highlighting risks spanning economic disruption, threats to public safety, and national security concerns. Project Glasswing is described as an urgent effort to direct AI’s capabilities toward protective and defensive applications.
Government Response and Ongoing Coordination
A Treasury Department spokesman underscored the administration’s commitment to preparing financial institutions for new cybersecurity challenges associated with AI advancements. “President Trump and the administration are continuing to engage on AI security in a thoughtful manner,” the official said. The White House is leading an interagency taskforce—including representatives from the Treasury—actively collaborating with government and industry to implement strategic plans that safeguard U.S. interests and citizens.
Further coordination efforts are planned, with Treasury scheduling ongoing meetings involving multiple regulators and financial entities to monitor developments related to AI risks, as well as other emerging issues.
The Federal Reserve declined to comment on the specifics of the meeting.
Broader Context on AI and Financial Stability
This high-level engagement comes amid heightened awareness of artificial intelligence as a systemic risk factor. In 2023, the Biden administration formally identified AI as a potential threat to financial stability—a unprecedented designation that spotlighted the need for vigilant oversight. The evolving dialogue reflects increasing recognition that AI technologies, while powerful tools, may also introduce complex cybersecurity challenges requiring proactive mitigation.
Authors: Aimee Picchi, Associate Managing Editor, MoneyWatch; Richard Escobedo, Economic Policy Correspondent
Edited by: Alain Sherter and Richard Escobedo
For more updates on AI, cybersecurity, and financial regulation, stay tuned to CBS News.