US appeals court lets Pentagon keep blocking Anthropic over national security risk designation
rohanpaul_ai · x · 2026-09-26
The D.C. Circuit ruled the Pentagon can continue excluding Anthropic from military contracts for now over its 'national security risk' designation, handing the Trump administration a win. The majority held Claude's built-in safety restrictions can qualify as a supply-chain risk because they blocked government-requested tasks and created uncertainty during an overseas operation — reasoning that turns on whether encoded safety rules can make the model refuse military tasks deemed operationally critical, not on remote control after deployment. The court accepted Anthropic's AI-safety advocacy as protected speech but said exclusion flowed from rejected contract terms, not retaliation. Judge Rita Lin's separate California ruling struck down the broader §3252 designation; both can coexist because §4713's supply-chain-risk definition is substantially broader, and the D.C. Circuit had 'no quarrel' with the California conclusion.
More from Safety
- Defending METR: public data can't second-guess its AI speedup findings — peterwildeford · 2026-11-11
- Australia's review of an OpenAI agent's unauthorized Medicare access exposes the agent accountability problem — cmb24k · 2026-11-11
- Adversarial RL envs for backdoor insertion vs. detection: where does it converge? — DimitrisPapail · 2026-11-11
- Using a third-party inference provider with your coding harness hands it shell access to your machine — TheZachMueller · 2026-11-11
- Democrats propose requiring federal permission to train AI models, with up to 20 years in prison — Polymarket · 2026-11-11
- WIRED asks 10 experts what the real AI nightmares are — not Skynet — AINowInstitute · 2026-11-11