David S. Kemp analyzes the first federal court ruling on AI and attorney-client privilege, United States v. Heppner, examining the court’s reasoning across each element of the privilege test and the work product doctrine. Mr. Kemp argues that while the court reached the correct result on two independent and sufficient grounds, its unnecessary confidentiality analysis was methodologically flawed—treating Anthropic’s broadest contractual reserved rights as dispositive without examining the specific terms, product tier, or training preferences that actually governed Heppner’s use. He warns that this overbroad reasoning, rather than the uncontroversial holdings, is what future courts will most likely cite, with potentially damaging consequences for privilege claims in any professional context involving third-party platforms.

Cornell Law professor Michael C. Dorf discusses the recent standoff between Anthropic and the Pentagon over Anthropic’s refusal to permit its AI tools to be used for mass surveillance or autonomous weapons, leading the Trump administration to designate Anthropic a national security supply-chain. Professor Dorf argues that while both mass surveillance and autonomous weapons deployment may already be unlawful under the Fourth Amendment and customary international law respectively, Anthropic had sound reasons to seek explicit contractual carveouts rather than rely on those legal limits—and that the Pentagon’s unwillingness to accept those carveouts raises the alarming inference that the administration intends to pursue both activities.