Bessent summons bank executives over Anthropic cyber risk

Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened a group of major bank executives at the Treasury Department in Washington, D.C., on Tuesday to address urgent cybersecurity risks emerging from the capabilities of Anthropic's newly released Mythos AI model. The meeting, confirmed by multiple sources to The Hill, was a proactive regulatory response to the potential for advanced artificial intelligence to be weaponized for sophisticated cyberattacks against the nation's critical financial infrastructure. The discussion focused specifically on the implications of Anthropic's latest large language model, which represents a significant leap in AI reasoning and code-generation abilities. Officials and executives analyzed hypothetical threat scenarios where such technology could be used to automate the discovery of software vulnerabilities, craft highly convincing phishing campaigns, or orchestrate complex, multi-vector attacks on banking networks. The gathering underscores a growing recognition within the highest levels of government that next-generation AI is not just a tool for economic innovation but also a potent new vector for systemic financial risk that requires coordinated defense. This high-level meeting reflects a shift in the regulatory posture of U.S. financial authorities, moving from general advisories on cyber hygiene to targeted, intelligence-driven briefings on specific technological threats. The involvement of both the Treasury and the Federal Reserve signals a whole-of-government approach to financial stability, treating AI-powered cyber threats with a seriousness previously reserved for traditional economic shocks. The dialogue is expected to inform new regulatory guidance and potentially catalyze increased public-private investment in AI-specific defensive technologies for the banking sector.