Treasury Secretary Bessent and Fed Chair Powell meet bank CEOs on Anthropic AI risks

U.S. Treasury Secretary Eleanor Bessent and Federal Reserve Chair Jerome Powell convened an emergency meeting with the chief executives of major U.S. banks in Washington, D.C., on Tuesday, following growing concerns about systemic financial risks posed by the rapid adoption of artificial intelligence systems developed by Anthropic. The closed-door session, which included leaders from JPMorgan Chase, Bank of America, Citigroup, and Wells Fargo, was called to assess potential vulnerabilities in the banking sector related to AI-driven decision-making, algorithmic trading, and cybersecurity threats. The meeting was prompted by a classified Treasury Department report, leaked last week, that highlighted several near-miss incidents where Anthropic's Claude AI models, used for credit scoring and fraud detection, exhibited unpredictable behavior that could have triggered cascading loan denials or erroneous trading signals. Regulators are particularly concerned about 'model monoculture,' where multiple major institutions rely on similar AI architectures from a single provider, creating a concentrated point of failure for the financial system. Discussions focused on establishing new regulatory frameworks for AI safety in finance, including proposed stress tests for AI models, mandatory disclosure of AI usage in core banking functions, and the creation of a 'circuit breaker' mechanism to halt AI-driven transactions during periods of market volatility. The bank CEOs reportedly expressed support for clearer guidelines but cautioned against overly restrictive regulations that could stifle innovation and put U.S. institutions at a competitive disadvantage globally. The outcome signals a significant shift toward proactive AI regulation in the financial sector, with the Treasury and Fed expected to release joint preliminary guidelines within 30 days. This meeting represents one of the highest-level governmental responses to AI risk since the White House's executive order on artificial intelligence last year, highlighting how financial stability has become a central concern in the AI governance debate.