March 10, 2026 — PM Anthropic Special Report
Anthropic, a public benefit corporation (PBC) operating the Claude AI, was in contract negotiations with the DOD when the department demanded full removal of Anthropic's usage policy — specifically: allowing Claude to surveil U.S. citizens and deploy lethal autonomous drone warfare without human oversight. Anthropic refused, citing its PBC mission (building a safer AI ecosystem), the fact that Claude's models make mistakes, and its own ethical guidelines. The President ordered all agencies to cease using Claude and cancel contracts with Anthropic. Secretary Hegseth accused Anthropic of "arrogance and betrayal" and designated it a "supply chain risk to national security." Anthropic filed a complaint and TRO request on March 9 alleging First and Fifth Amendment violations and APA violations. Google and OpenAI employees filed personal amicus briefs arguing that unfettered AI access without human oversight was dangerous to the public interest. Bryan covered the March 10 status hearing (Judge Lin, San Francisco, available via Zoom) live, noting this might be the only public hearing given the DOD's likely move to close future proceedings.
No new substantive update in this episode; the case was mentioned in the context of oral argument week wrap-up. See 0304_SCOTUS.md for full entry.