WASHINGTON, April 13 - Anthropic is holding discussions with officials in the U.S. government about its latest AI model, Mythos, and intends to brief authorities on future models as well, the company’s co-founder said on Monday. The outreach comes despite a recent break in business relations with the Pentagon after a disagreement over contract terms and use restrictions.
Anthropic co-founder Jack Clark spoke at the Semafor World Economy event in Washington and said the company remains engaged with the federal government. He described the disagreement with the Department of Defense as a "narrow contracting dispute," and emphasized Anthropic’s interest in national security concerns. "We have a narrow contracting dispute, but I don’t want that to get in the way of the fact that we care deeply about national security," he said.
Clark added that Anthropic believes the government should be informed about advanced AI systems, and that the company is communicating about Mythos and plans for subsequent models. "Our position is the government has to know about this stuff ... So absolutely, we’re talking to them about Mythos, and we’ll talk to them about the next models as well," he said. The specific agencies involved and the details of those conversations were not made clear.
Mythos was announced on April 7 and Anthropic described it in a company blog post as its most capable model yet for coding and agentic tasks, highlighting the system’s ability to act with autonomy. Experts have cautioned that Mythos’ advanced coding capabilities could give it a heightened capacity to detect cybersecurity weaknesses and to devise methods that might exploit them.
Last month the Pentagon labeled Anthropic a supply-chain risk, blocking the company’s tools from being used by the department and its contractors following the contract dispute over guardrails for military applications of the company’s artificial intelligence. A Washington, D.C., federal appeals court last week declined to block the Pentagon’s national security blacklisting for the time being, a result the Trump administration viewed as favorable. That decision followed an earlier appeals court ruling that reached an opposite conclusion in a separate legal challenge brought by Anthropic.
Investment and data advisory content included in the original report was also present at the event. The promotional material argued that better data improves investment decisions and mentioned a product that combines institutional data with AI-powered insights to aid in identifying investment opportunities. It posed a question about the best investments of 2026 and referenced an AI-driven tool for investors.
The company’s public statements and the judicial developments underscore continuing uncertainty about how advanced AI systems will be governed when used in defense and related sectors, while Anthropic seeks to maintain dialogue with U.S. officials about its technology and next-generation models.