Anthropic briefs Trump administration on Mythos AI model despite ongoing lawsuit against DoD

Anthropic co-founder Jack Clark has confirmed that the company briefed the Trump administration on its powerful new AI model, Mythos, despite an ongoing lawsuit against the U.S. Department of Defense. Announced last week, Mythos possesses advanced cybersecurity capabilities—such as finding and exploiting unknown software flaws—that Anthropic deems too dangerous for public release, limiting access to select tech and cyber organizations for testing.

In an interview at Semafor’s World Economy summit this week, Clark, who also serves as Anthropic’s Head of Public Benefit, emphasized the need for close government collaboration on such revolutionary technologies. “Our position is the government has to know about this stuff, and we have to find new ways for the government to partner with a private sector that is making things that are truly revolutionizing the economy, but are going to have aspects to them which hit national security equities,” Clark said, as reported by TechCrunch. He added that Anthropic would continue briefing officials on Mythos and future models.

This disclosure highlights a complex dual-track strategy by Anthropic: engaging with the government on AI risks while litigating against it. In March, the company sued the DOD after the agency labeled Anthropic a supply-chain risk, amid disputes over military access to its systems. Clark’s comments underscore the tension, as the firm navigates national security concerns without fully severing ties.

Reports indicate Trump administration officials have encouraged major banks—including JPMorgan Chase, Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley—to test Mythos. Federal agencies are also finding ways to evaluate the model, with Treasury Department IT officials seeking to use it for patching network vulnerabilities, according to Bloomberg. An anonymous Anthropic official confirmed briefings with U.S. government officials on Mythos’s hacking abilities, though no agencies were named in the company’s announcement.

The limited rollout of Mythos has drawn international praise. Canada’s AI minister lauded Anthropic’s cautious approach of introducing the model to select companies first, as noted in Bloomberg coverage. This reflects broader industry debates on balancing AI innovation with safety.

For those affected—ranging from financial institutions bolstering cyber defenses to government agencies addressing vulnerabilities—Mythos represents a double-edged tool. Its capabilities could strengthen national cybersecurity but raise fears of misuse in attacks or espionage.

Looking ahead, Anthropic’s ongoing dialogues signal a push for new public-private partnerships in AI governance. Clark also touched on wider societal impacts, like potential unemployment and shifts in higher education, urging proactive policy responses to AI’s transformative power.