Bessent, Powell warned bank CEOs about Anthropic model risks

US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell recently met with CEOs of major banks to raise concerns about cybersecurity risks linked to Anthropic’s latest AI model, known as Claude Mythos. This model has elevated alarm over its ability to detect and exploit vulnerabilities across financial systems at unprecedented speed and scale, posing potential threats to the stability and security of the financial sector. The meeting underscored the need for banks and regulators to exercise caution when adopting advanced AI tools, as the risks extend beyond individual institutions to systemic financial threats. Anthropic’s model compresses the timeline from vulnerability discovery to exploitation, altering traditional risk frameworks and challenging existing security measures. Experts describe the AI capability as “exceptionally dangerous” for its potential to reveal previously hidden weaknesses that standard scanners may not detect. This development has prompted several international financial bodies, including Canadian regulators, to engage in dialogue about enhancing cyber resilience and governance in AI-driven environments. These discussions reflect a growing recognition that while AI offers transformative opportunities in finance, it also introduces novel risks that require vigilant oversight, robust data governance, and comprehensive scenario planning. The proactive engagement by Treasury and the Fed highlights the broader government strategy to balance AI innovation with safeguarding critical financial infrastructure against emerging cyber threats.