Federal Reserve Chair Jerome Powell and U.S. Treasury Secretary Scott Bessent met with major U.S. bank CEOs this week to address potential cybersecurity risks linked to a new AI model from Anthropic. The meeting, held in Washington, D.C., was convened specifically to discuss concerns about the model, known as Mythos, according to a source familiar with the discussions.
Anthropic recently released Mythos in a limited capacity, citing concerns that its capabilities could be exploited by malicious actors, including hackers. While details about the model remain scarce, the decision to restrict access reflects increasing caution among AI developers as systems grow more powerful and potentially dual-use.
The involvement of top financial regulators and banking leaders signals rising concern about how advanced AI tools could impact critical infrastructure, particularly in the financial sector. Banks are frequent targets of cyberattacks, and more capable AI systems could potentially accelerate the sophistication of such threats.
The meeting underscores a broader shift in how policymakers are approaching AI risk, moving from theoretical concerns to direct engagement with industry stakeholders. As AI capabilities advance, regulators and companies alike are facing mounting pressure to balance innovation with safeguards that prevent misuse, especially in sectors where security and stability are paramount.
