The secretary of, Scott Bessantbrought together executives from large banks to discuss cyber risks associated with new artificial intelligence models. The meeting was held in Washington at the beginning of the week and was attended by Jerome Powellpresident of .
Representatives of Bank of America, Citigroup, Goldman Sachs, Morgan Stanley e Wells Fargo were at the meeting. O CEO do , Jamie Dimonwas invited, but did not attend.
The meeting was motivated by concerns about the Claude Mythos model, developed by Anthropic. According to authorities, the system has an advanced ability to identify vulnerabilities in critical software, which can be exploited by malicious agents.
Anthropic launched the Claude Mythos Preview on Tuesday (Apr 7, 2026) to a select group of partners, including Amazon, Apple e Microsoft. According to the company, the model has already identified thousands of serious flaws, some of which have existed for decades.
Financial sector executives highlighted the growing risk. In an annual letter, Dimon stated that cyber threats remain among the main challenges and tend to intensify with the advancement of AI.
Anthropic reported that it is in dialogue with the United States government about the model’s capabilities. Neither banks nor authorities officially commented on the meeting.
