White House officials reviewing Pentagon’s decision to remove anthropic risk label from AI systems, emphasizing national secu

Editorial illustration for White House drafts action to lift Pentagon’s Anthropic risk label

White House drafts action to lift Pentagon’s Anthropic...

White House drafts action to lift Pentagon’s Anthropic risk label

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

May 3, 2026 • Updated: May 7, 2026 • 2 min read

In recent weeks the White House has drafted a measure that would effectively override the Defense Department’s current risk label on Anthropic. The move follows a high‑level meeting that brought together the administration’s chief of staff, Susie Wiles, Treasury Secretary Scott Bessent, and Anthropic founder‑CEO Dario Amodei. The friction started after Anthropic declined to sign the Pentagon’s standard contract, which would have permitted its Claude system to be used for any activity deemed lawful.

Instead, the company demanded explicit exclusions for large‑scale internal monitoring and for fully autonomous weapon platforms. Competitors such as OpenAI and Google have already entered into the government’s standard agreements without similar carve‑outs. The draft action signals that officials are willing to adjust the procurement framework in order to keep Anthropic’s technology within the federal supply chain, even as the firm holds out for tighter usage restrictions.

A draft executive action could sidestep the Pentagon's supply chain risk designation of Anthropic. One source described the push as a way to "save face and bring em back in." Earlier this month, White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent sat down with Anthropic CEO Dario Amodei. The dispute began when Anthropic refused to sign a Pentagon agreement that would have allowed the use of its Claude model for "all lawful purposes." The company insisted on carving out bans on "mass domestic surveillance" and "fully autonomous weapons." OpenAI and Google have already signed similar deals with the Pentagon.

White House moves to restore Anthropic access after Pentagon standoff - THE DECODER

Why this matters

The draft guidance signals a clear shift. By allowing agencies to tap Anthropic’s Mythos model, the administration appears ready to move past the Pentagon’s supply‑chain risk label. Yet the executive action would effectively sidestep that designation, raising questions about how risk assessments will be reconciled across departments.

One source framed the effort as a bid to “save face and bring ’em back in,” suggesting political calculus as much as technical need. Earlier this month, White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent met with Anthropic’s CEO, a detail that hints at high‑level coordination but offers no insight into the criteria that will govern future access. Whether the new policy will satisfy both security concerns and the agency’s AI ambitions remains uncertain.

The administration’s next steps will likely clarify how the balance between risk mitigation and operational capability is intended to play out, but for now the exact parameters of the restored relationship are still unclear.