Statement from Dario Amodei on our discussions with the Department of War

• Anthropic refuses DoD’s “any lawful use” demand; two hard limits held.
  • DoD threatened supply chain risk designation and removal from contracts.
• Red line 1: mass domestic surveillance contradicts democratic principles.
  • AI can stitch scattered innocuous data into full life portraits, at scale.
• Red line 2: fully autonomous weapons — AI not reliable enough yet.
  • Proper guardrails don’t exist today; open to revisiting if they do.
• First frontier AI company deployed across classified networks + NSA-tier agencies.
• Forfeited “several hundred million dollars” blocking CCP-linked entity access.
• Amodei: willing to support transition to alternative provider if required.

X discourse

• @DropSiteNews: “The Untold Story of Dario Amodei’s Antiwar Past, as His AI Tech is Used for a New Middle East Wa” (240 likes)
• @ColinPClarke: “This is not a conspiracy; this is an assessment of the Administration’s own language & statements” (265 likes)
• @USWREMichael: “@SarahKHeck begged to see how @AnthropicAI could fix their behavior” (354 likes)
• @DavidSacks: “The Effective Altruist movement has a structural problem when it comes to conservative America” (5519 likes)
• @rsayre: “The DoD response didn’t deny reporting, but said it was ‘highly exaggerated and distorted’” (10 likes)
• @sfrantzman: “‘In light of grossly false and distorted recent reporting, the Department of War repeats its statement’” (2 likes)

Dario Amodei, CEO Anthropic (co-founder, ex-OpenAI VP Research) · ** · Read on anthropic.com