The AI Policy Podcast Anthropic Goes to Court While Claude Goes to War in Iran
36 snips
Mar 11, 2026 
Gregory C. Allen, senior adviser at Wadhwani AI Centers and AI policy and national security expert, explains the Anthropic-Pentagon legal clash. He traces the lawsuits and why venue matters. He then shifts to how Claude and government models are being used in Iran, covering intelligence acceleration, targeting workflows, and risks to data-center infrastructure.
AI Snips
Chapters
Transcript
Episode notes
Supply Chain Law Forces Classified One-Way Court Review
- The supply chain risk statute (41 U.S.C. §4713) limits judicial review to the D.C. Circuit to protect classified evidence and methods.
- Gregory C. Allen explains this creates one-way classified submissions the company cannot review, complicating Anthropic's legal strategy.
Use Public Statements To Establish Broader Court Standing
- To avoid losing broader review rights, Anthropic argues harms from tweets and public statements give standing outside §4713's exclusive path.
- They seek a California preliminary injunction to block harms from Truth Social and Hegseth tweets.
Court Filings Reveal Pentagon Praise And Dependence
- Anthropic filed sworn court evidence detailing praise from Pentagon officials to show reliance on Claude and contradictory government behavior.
- Court filings quote Hegseth calling Anthropic's capabilities "exquisite" and officials saying "we need them and we need them now."