DC Circuit lets Trump keep Anthropic blacklisted from federal, defense use
A divided appeals court says the administration can bar Anthropic's Claude over withheld military features, even absent malicious intent.
What to know
- A divided DC Circuit panel (2-1, both judges Trump appointees) upheld the administration's authority to blacklist Anthropic's Claude for withheld military features.
- The ruling relies on a different, more permissive statute (41 U.S.C. § 4713) than the one a district court used last month to find the same blacklisting illegal.
- Anthropic says it is considering further review, potentially en banc rehearing or a Supreme Court petition.
- Despite the legal fight, Commerce Secretary Howard Lutnick has said the administration and Anthropic are 'in tune.'
“Anthropic raises the deeply sobering prospect of unconstrained AI models hallucinating inappropriate targets for lethal military force.”
US Court of Appeals for the DC Circuit, appellate court panel · Ars Technica ↗ · Sep 24
Anthropic AI company challenging the blacklisting
Donald Trump US President
Pete Hegseth US Defense SecretaryJudge Gregory Katsas DC Circuit appeals judge
Judge Neomi Rao DC Circuit appeals judge
Howard Lutnick US Commerce Secretary
How it unfolded 2 developments, newest first · click a bar or a number to jump articlesposts
-
2
Anthropic says it will consider further legal options
An Anthropic spokesperson said the company disagrees with the ruling and noted a separate court had already found the government's parallel designation unlawful.
“We respectfully disagree with the court's decision. Another federal court has already held the government's parallel designation unlawful. We remain confident in our position and are considering all options, including further review.”
— Anthropic spokesperson -
S
Pentagon judges ruled that "overly constrained AI models" risk failing military operations, allowing the department to… #Semiconductors #AIInfrastructure #GPU #DefenseTech
2 more of the top 3 · 5 posts in this stretch
-
Pointing out the categorical difference is correct. But a military fails as soon as it is surpassed by enemy capabilities in conflict, meaning the threshold of failure increases with technological advancement. Having two powers, one adopting AI-assisted systems, and the other not, makes the latter quickly lose, even though it hasn't lost any…
-
S
Court rules Pentagon can blacklist Anthropic for refusing to enable Claude features - Ars Technica
-
-
1
DC Circuit rules 2-1 that Trump can blacklist Anthropic
The appeals court held that under a separate, more permissive statute (41 U.S.C. § 4713) over which it has exclusive jurisdiction, the Defense Secretary did not exceed his authority in blacklisting Anthropic even without malicious intent, denying Anthropic's petition for review.
“the Secretary did not transgress any limits on his authority under the Supply Chain Security Act or the Constitution. Accordingly, we deny the petitions for review.”
— US Court of Appeals for the DC Circuit -
1 outlet first by Ars Technica, 5d ago · read ↗
-
-
background
District court rules the blacklisting illegal — A judge in the US District Court for the Northern District of California found the action unlawful because Anthropic does not meet the legal definition of a supply-chain risk, which is limited to malicious actions by adversaries under 10 U.S.C. § 3252.
-
background
DC Circuit denies Anthropic's emergency stay request — The same appeals court that later ruled against Anthropic first denied its emergency motion for a stay of the blacklisting.
-
background
Anthropic sues Trump administration over blacklisting order — Anthropic sued after Trump and Defense Secretary Pete Hegseth ordered federal agencies to stop using its products and banned defense contractors from doing business with the company.