Federal Reserve Chair Jerome Powell, Treasury’s Scott Bessent, and Top Bank CEOs Convene to Address Cybersecurity Risks of Anthropic’s AI Model Mythos
April 10, 2026 | CBS News
In a rare closed-door meeting held on Tuesday, Federal Reserve Chair Jerome Powell and Treasury Secretary Scott Bessent gathered with chief executives from leading banks to discuss the cybersecurity implications posed by Anthropic’s latest advanced artificial intelligence (AI) model, Mythos. Sources familiar with the discussions confirmed the meeting to CBS News.
The meeting took place at the Treasury Department’s headquarters in Washington, D.C. and aimed to address the emerging risks associated with new AI technologies capable of identifying vulnerabilities in major operating systems and web browsers.
Attendance and Context
While invited, JPMorgan Chase CEO Jamie Dimon was unable to attend. Nonetheless, the gathering assembled top financial industry leaders alongside key federal officials to appraise the potential threats to financial infrastructure that could arise from the rapid proliferation of powerful AI systems like Mythos.
Bloomberg News had earlier reported the meeting, noting Powell and Bessent’s intent to engage financial executives about the evolving security landscape. The concern extends beyond banking to broader public safety, economic stability, and national security considerations.
About Anthropic’s Mythos Model
Anthropic, the AI research company known for its generative chatbot Claude, recently unveiled Mythos — a cutting-edge AI model with the capability to uncover hidden cybersecurity weaknesses. Given its advanced functionality, Anthropic announced it would avoid releasing Mythos widely to prevent misuse.
The company highlighted the urgency of responsibly harnessing Mythos’s capabilities through a collaborative effort with major technology firms, including Amazon, Apple, and Nvidia. This initiative, dubbed Project Glasswing, is focused on deploying Mythos to bolster cybersecurity defenses.
“Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely,” Anthropic stated in a recent post. The company warned that unchecked dissemination could lead to severe consequences affecting economies, public safety, and national security.
Government Response and Ongoing Coordination
A Treasury Department spokesperson emphasized the federal government’s proactive posture in addressing AI-related security challenges. “The White House has been leading an ongoing core interagency taskforce, which includes the Treasury, that has been proactively engaging across the government and industry to execute the first phases of a plan to ensure the United States and Americans are protected,” the spokesperson said.
The spokesman also noted that additional coordination meetings involving various regulators and financial institutions are planned as part of an ongoing strategy to manage emerging AI developments alongside other critical issues.
Federal Reserve officials declined to comment directly on the meeting.
Background on AI and Financial Stability
Concerns over the risks posed by AI to financial stability are not new. In 2023, the Biden Administration formally recognized AI as a potential threat to the financial system’s stability—an unprecedented designation underscoring the urgency of addressing these emerging technologies’ risks.
Looking Ahead
As AI models evolve rapidly, financial and government leaders face the complex challenge of balancing innovation with security. The collaboration between Anthropic, tech giants, and government entities reflects a wider effort to utilize AI defensively, aiming to protect critical infrastructure while mitigating the dangers of misuse.
The discussions at this high-level meeting represent a key step toward fostering a secure framework for AI’s integration into sectors vital to the nation’s economic wellbeing.
Reporters: Aimee Picchi and Richard Escobedo, CBS News
Edited by Alain Sherter and Richard Escobedo
© 2026 CBS Interactive Inc. All rights reserved.