Anthropic briefed Trump administration on Mythos model despite Pentagon lawsuit

Anthropic co-founder Jack Clark confirmed the company briefed the Trump administration on Mythos, a model the company announced last week but is withholding from public release due to what it describes as dangerous cybersecurity capabilities.

Speaking at Semafor's World Economy summit this week, Clark explained Anthropic's position on government engagement despite the company's March 2025 lawsuit against Trump's Department of Defense. The DOD labeled Anthropic a supply chain risk after the company refused to provide unrestricted access to its AI systems for use cases including mass surveillance of Americans and fully autonomous weapons. OpenAI won that contract instead.

Government briefings continue

"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, and other ones," Clark said. "So absolutely, we talked to them about Mythos, and we'll talk to them about the next models as well."

Clark characterized the DOD labeling as a "narrow contracting dispute" that shouldn't interfere with Anthropic's commitment to national security.

According to reports last week, Trump officials encouraged major banks including JPMorgan Chase, Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley to test Mythos.

Employment impact assessments

Clark, who leads Anthropic's team of economists, diverged from CEO Dario Amodei's previous warnings that AI advances could bring Depression-era unemployment levels. Clark said Anthropic is currently seeing "some potential weakness in early graduate employment" across select industries, though the company is prepared for major employment shifts.

On education, Clark suggested students pursue majors "that involve synthesis across a whole variety of subjects and analytical thinking." He explained that AI provides access to "an arbitrary amount of subject matter experts in different domains," making cross-disciplinary thinking more valuable than domain expertise alone.

What this means

Anthropic is navigating a complex position: suing the Pentagon over military AI use while simultaneously briefing the administration on models it considers too dangerous for public release. The company's approach suggests it views government awareness of frontier AI capabilities as a national security imperative, even when commercial and defense relationships are strained. The lack of public information about Mythos's specific capabilities or release timeline leaves significant questions about what threat model justified both withholding the model and briefing government officials.