Anthropic co-founder confirms the company briefed the Trump administration on Mythos | TechCrunch
Jack Clark, Anthropic’s co-founder and head of public benefit, confirmed the company briefed the Trump administration on its withheld Mythos model due to its powerful cybersecurity capabilities. He downplayed the Pentagon’s “supply-chain risk” label while defending continued government engagement and also discussed AI’s potential impact on jobs and higher education.
- Anthropic confirmed it briefed the Trump administration on Mythos, an unreleased AI model withheld due to its powerful cybersecurity capabilities, and will keep briefing officials on future models.
- Trump officials reportedly pushed major banks (JPMorgan, Goldman Sachs, Citigroup, Bank of America, Morgan Stanley) to test Mythos, which Clark confirmed.
- Anthropic is simultaneously suing the DOD over a "supply-chain risk" label tied to a contract for mass surveillance/autonomous weapons work that OpenAI won, which Clark calls a narrow contracting dispute rather than a sign of broken government ties.
- Clark downplayed CEO Dario Amodei's warnings of Depression-era job losses from AI, citing only early signs of weak graduate employment in a few industries.
Anthropic co-founder Jack Clark confirmed at the Semafor World Economy summit that his team briefed the Trump administration on Mythos, the company’s latest AI model. Mythos isn’t public because of its potent cybersecurity capabilities. Clark said Anthropic wants the government looped in on any technology that could affect national security, even as it’s suing the Department of Defense over being labeled a supply-chain risk. That dispute centers on Pentagon demands for unrestricted access to Anthropic’s systems for projects like mass surveillance and autonomous weapons—a contract OpenAI ultimately won.
Reports last week suggested Trump officials urged banks—including JPMorgan Chase, Goldman Sachs, Citigroup, Bank of America and Morgan Stanley—to test Mythos. Clark confirmed those conversations and said Anthropic will brief officials on future models too. He framed the DOD lawsuit as “a narrow contracting dispute” that shouldn’t obscure the broader goal: building partnerships between the private sector and government to handle AI’s security implications responsibly.
On AI’s social impact, Clark clashed slightly with Anthropic CEO Dario Amodei’s dire unemployment projections. While Amodei warns of Depression-era job losses, Clark’s team of economists sees only “some potential weakness in early graduate employment” in a few industries so far. He urged students to study fields that demand “analytical thinking” and “synthesis across a whole variety of subjects,” since mastering the right questions and combining insights across domains will matter more than memorizing facts.