Trump officials may be encouraging banks to test Anthropic’s Mythos model

Top officials in the Trump administration are reportedly encouraging major U.S. banks to adopt Anthropic’s new Mythos AI model for cybersecurity, a surprising development given the administration's ongoing legal conflict with the AI company. According to a Bloomberg report, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell recently met with bank executives to promote the use of Mythos for detecting financial system vulnerabilities. This move creates a striking contrast with the Department of Defense's designation of Anthropic as a supply-chain risk, a label that stemmed from a dispute over government use of its AI.

Anthropic announced Mythos this week, noting that access would be limited due to its potent, albeit unintentional, capabilities in discovering security flaws. While JPMorgan Chase was the only bank publicly named as an initial partner, reports indicate that Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley are also testing the model. The company's claims about Mythos's power have been met with some industry skepticism, with observers suggesting it could be a well-crafted enterprise sales strategy rather than a reflection of unique technical capabilities. Mythos was not specifically trained for cybersecurity tasks, which makes its reported effectiveness in the area a subject of intense interest.

This situation highlights a growing and complex tension in how governments engage with advanced AI systems. While one arm of the administration views Anthropic with suspicion over control and safety protocols, another sees its technology as a vital tool for securing critical national infrastructure. The dual posture—treating a company as both a partner in national security and a potential risk—underscores the challenge of regulating powerful, general-purpose AI. The discussion is not limited to the U.S., as financial regulators in the United Kingdom are also reportedly analyzing the potential risks introduced by the Mythos model.

The administration's simultaneous legal challenge and promotion of Anthropic's Mythos model exposes a fundamental government conflict: how to treat powerful AI as both a strategic national asset and a potential systemic risk.