White House Moves to Reverse Anthropic AI Restrictions – Fact Check

Claim 1 The White House is developing a plan to allow federal agencies to use Anthropic’s newest AI models while bypassing the supply‑chain risk designation previously applied to the startup.

Evidence Axios reported on April 29 that unnamed sources said the administration is drafting guidance that would let agencies get around the designation and onboard models including the cyber‑focused Mythos system. Nextgov also noted the White House is crafting similar guidance to clear the way for government use of Anthropic tools.

Verdict Mostly true.

Analysis Multiple outlets confirm the White House is working on guidance to ease the restriction, but the plan remains in development and could change before final issuance.

Claim 2 In late February the White House instructed federal agencies to cease using Anthropic’s AI products.

Evidence Axios and Nextgov both reported that the White House told agencies in late February to stop using Anthropic’s technology, citing a dispute over domestic surveillance and autonomous weapons use.

Verdict True.

Analysis The instruction is directly attested by two independent news reports, confirming the White House’s directive to halt agency use of Anthropic’s AI.

Claim 3 Former President Donald Trump announced that the federal government would stop working with Anthropic and that agencies using Anthropic’s Claude models would have a six‑month phase‑out period.

Evidence The White House’s February directive originated from the Trump administration, not from Trump personally, and no official announcement specified a six‑month timeline for Claude models.

Verdict False.

Analysis The evidence shows the decision was administrative, not a personal statement by Trump, and the six‑month period lacks corroboration.

What to watch next Monitor whether the White House finalizes its guidance to allow agency use of Anthropic’s models and how any legal challenges from the company evolve.