On April 7, 2026, U.S. Treasury Secretary Bessent and Federal Reserve Chair Powell held a meeting with top executives from systemically important banks on Wall Street. During this gathering, they advocated for the utilization of Anthropic’s newly released Mythos model to uncover potential vulnerabilities within the financial system. Although Mythos was not explicitly designed for cybersecurity purposes, it has demonstrated exceptional proficiency in identifying system weaknesses.

Initially, access to the Mythos model was restricted to a select group of companies for testing, including tech behemoths like Amazon and Google, as well as financial institutions such as JPMorgan Chase. Presently, Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley are also conducting internal tests, with broader access anticipated in the near future.

Previously, the U.S. Department of Defense had classified Anthropic as a “supply chain risk,” a designation that sparked legal disputes between the government and the company. The government’s current initiative to encourage banks to adopt the Mythos model underscores a recognition of AI’s value, while also revealing a somewhat contradictory stance.