A Federal Appeals Court Says the Pentagon Can Call Anthropic a National Security Risk for Building an AI That Refuses to Spy on Americans
About This Episode
The DC Circuit ruled 2-1 on Friday that the Defense Department had ample grounds to designate Anthropic a supply chain risk because the company encodes restrictions into Claude, including refusals on fully autonomous lethal weapons and mass surveillance of Americans, that could stop the model performing tasks the Pentagon deems necessary. The label keeps Claude out of Defense Department systems and contractor work; Anthropic says it has cost billions in lost business ahead of its IPO and is weighing further review. The dissent warns that any contractor can now be told to change its AI-use policy or be branded a national security threat.
Our Take
A court just decided that an AI company's conscience, encoded as refusals, is a national-security risk when the government wants the tool to do what the vendor won't; the same week the founders asked to keep that conscience in their own hands after the IPO, so the episode asks who gets to decide what an AI will not do.
Continue Reading on Unscarcity
AI Governance: Why Hawk-Eye Works But Doesn't Make Rules
The framework separates the machine that enforces from the humans who decide what is worth enforcing; this ruling is a fight over which humans get to be the conscience wired into the tool, the vendor that built it or the secretary who buys it.
Mill's 1859 Harm Principle Meets 2026 AI Algorithms
Pettit's distinction between interference and domination fits both the surveillance restriction Anthropic kept and the dissent's warning that every future contractor now lives at the secretary's pleasure.