FOLLOW US

U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell held an urgent meeting this week with top bank executives to flag potential cybersecurity threats tied to a new artificial intelligence system developed by Anthropic, according to two people familiar with the discussions.

The company earlier this week introduced its advanced Mythos model but refrained from rolling it out widely, pointing to concerns that it could reveal previously undiscovered security flaws.

Anthropic has said the system is capable of identifying ⁠and exploiting weaknesses, opens new tab across "every major operating system and every major web browser".

Last week, Anthropic said it was in ongoing discussions with U.S. government officials about the model's "offensive and defensive cyber capabilities."

A third person familiar with the situation confirmed the company’s outreach, noting that Anthropic had already briefed senior U.S. officials and leading industry players on Mythos before unveiling it.

The meeting, hosted by the Treasury Department in Washington on Tuesday, was intended to ensure major banks understand the risks associated with Mythos and comparable AI tools, and are taking precautions to safeguard their systems, one of the sources said.

Invitations went out as many chief executives from the largest U.S. banks were already in Washington for previously scheduled engagements, one source added.

Anthropic has indicated that access to Mythos will initially be restricted to roughly 40 technology firms, including Microsoft and Google.

According to Bloomberg News, executives from Citigroup, Morgan Stanley, Bank of America, Wells Fargo and Goldman Sachs attended the session. Jamie Dimon, chief executive of JPMorgan Chase, was unable to participate, one of the sources said.

Goldman Sachs, Wells Fargo and the Federal Reserve declined to comment, while the Treasury Department, the banks involved and Anthropic did not immediately respond to requests for comment.