US government to pre‑review AI models as Anthropic dispute eases

President Donald Trump signed a decree that lets the United States government review advanced artificial‑intelligence models before they are released. Major AI firms—including Google, OpenAI and Anthropic—will voluntarily submit their models for a 30‑day evaluation by a coordination center formed by the Treasury, the Department of Homeland Security and CISA. The decree aims to strengthen digital security, protect intellectual property and address vulnerabilities shown by Anthropic’s “Mythos” model.

Earlier this year Anthropic was placed on a Department of Defense supply‑chain risk list after refusing to allow U.S. forces to use its models for domestic surveillance and autonomous weapons. The company contested the designation in court. Relations improved after Anthropic CEO Dario Amodei met with White House officials in April, and the company announced plans for an initial public offering. "Whenever the government signals it is washing its hands of a company, that’s a big problem for that company," said attorney Franklin Turner.

Anthropic said it expects to collaborate on implementing the new AI decree, while the DoD continues to defend its stance on the supply‑chain risk ruling.