Treasury Secretary Bessent Leads AI Policy Push After Banks Flag Risks to Fed Payment Systems

Treasury Secretary Scott Bessent has taken a more active role in artificial intelligence policy this year after financial institutions warned that advanced models could expose critical infrastructure, including Federal Reserve payment rails, to outage risks, Semafor reported.

Bessent met with then-Fed chair Jerome Powell and CEOs of systemically important banks in April to discuss threats posed by Anthropic’s Mythos model. Banks have worked with the Treasury Department and Fed for years on backup transfer methods in case of a major disruption.

This spring, models such as Mythos increased pressure on those preparations. ” He described himself as “one of the point people on our AI policy” and said officials are balancing “innovation and safety” to maintain a lead over China. An executive order established a voluntary review process for advanced AI models and placed Treasury in charge of a new cyber clearinghouse.

The Trump administration separately asked OpenAI to limit release of a new advanced model over national security concerns. Anthropic is now in talks with the administration on restoring access to Mythos and its consumer version Fable after Commerce Department export controls took effect. Those discussions could slow the executive order’s rollout.

Federally and state-chartered institutions with master accounts at a Federal Reserve bank can use systems such as Fedwire for instant transfers. Fedwire handled more than 54 million transfers totaling over $283 trillion in the first quarter of 2026, averaging 886,695 transfers daily at more than $5 million each.