Fed Chairman Powell, Treasury's Bessent Discussed Anthropic AI Cyber Threat with Major U.S. Banks

Federal Reserve Chairman Jerome Powell and Treasury Secretary Scott Bessent convened a meeting with major U.S. bank chief executives to discuss potential cybersecurity threats from Anthropic's new Mythos AI model. The meeting addressed the implications of advanced AI systems capable of identifying and exploiting software vulnerabilities at scale. Anthropic rolled out its Mythos AI model to a select group of companies over concerns that hackers could exploit its capabilities for cyberattacks. Anthropic says the model, which it has designated for controlled deployment through its Project Glasswing initiative, demonstrated significant advances in identifying security flaws in code. The AI company is providing access to 40 technology companies along with $100 million in usage credits to help assess and mitigate potential risks. Participants in the government meeting included major financial institutions whose systems could be targets of AI-enabled cyberattacks. The discussions focused on preparing the banking sector for a new generation of AI-powered cybersecurity threats. The meeting reflects growing government concern about the dual-use nature of advanced AI systems. While capable of helping organizations identify and patch security vulnerabilities, these tools could also enable sophisticated attacks against critical infrastructure. Anthropic has positioned Project Glasswing as a responsible mechanism for distributing powerful AI capabilities while maintaining safeguards. The company says it is working with government officials and industry partners to understand and mitigate potential risks. Neither the Federal Reserve nor the Treasury Department released official statements about the meeting's outcomes.