Treasury Secretary Scott Bessent and Federal Reserve Chair Jay Powell summoned top Wall Street executives to discuss cybersecurity risks posed by advanced artificial intelligence models like Anthropic's Claude Mythos Preview. The emergency meeting, held Tuesday with bank CEOs already in Washington, focused on safeguarding against potential vulnerabilities as AI capabilities expand. Bank of America, Citigroup, Goldman Sachs, Morgan Stanley, and Wells Fargo executives attended the closed-door session.

The meeting followed Anthropic's announcement restricting its latest AI model to select companies for identifying critical cybersecurity flaws. While some critics dismissed the move as marketing, Powell and Bessent's actions underscore growing regulatory concerns about AI's potential to democratize sophisticated cyberattack capabilities. JPMorgan Chase CEO Jamie Dimon, who couldn't attend, had already acknowledged in his shareholder letter that AI could create "cybersecurity vulnerabilities."

Corporate leaders and policy experts have warned about increasingly capable AI models. A former Microsoft executive told The Times that advanced systems like Mythos could enable more widespread cyber threats. The Treasury and Fed officials' proactive approach reflects mounting anxiety about AI's potential to disrupt financial stability, particularly as banks face pressure to adopt new technologies while maintaining robust security protocols.