Anthropic applies for en banc rehearing by the U.S. Court of Appeals over the national security risk designation

Article

Anthropic for insisting that Claude the safety red line be classified as a national security risk, 2 than 1 applies for rehearing after losing the case Anthropic The lawsuit against the Pentagon is not over yet.October 13，Anthropic has asked the full U.S. Court of Appeals for the District of Columbia Circuit to rehear the blacklist case.September 25, the three-judge panel previously ruled by 2 versus 1 in support of the Pentagon, allowing Anthropic to remain listed as a national security supply-chain risk company.Anthropic said the ban had already cost it tens of billions of USD in business.

The two sides were originally partners. 2025 In July，Anthropic secured a contract worth 200000000 USD from the Pentagon, but the deal later fell apart over Claude's restrictions on its use.Anthropic insists on two bottom lines:AI must not be used for mass surveillance of Americans, nor may AI autonomously decide to launch lethal attacks. The military, meanwhile, demands the ability to use Claude for all lawful military purposes. This year,March the Pentagon placed Anthropic on the blacklist, prohibiting the continued use of its technology in defense projects.

The dispute is this: if safety restrictions are imposed on AI why would that instead constitute a national security risk? The Pentagon argues that Claude had previously refused to carry out government tasks because of built-in restrictions. If it were suddenly unable to respond to orders during wartime, military operations could be affected. Two judges accepted this reasoning.Anthropic countered that the relevant law was originally intended to prevent malicious disruption of supply chains, but has now been expanded to cover ordinary model-safety restrictions set by companies.

Original link https://m.theblockbeats.info/flash/371772

Key points

October 13，Anthropic has asked the full U.S. Court of Appeals for the District of Columbia Circuit to rehear the related blacklist case.
September 25, the three-judge panel ruled 2 versus 1 in support of the Pentagon, allowing Anthropic to remain listed as a national security supply-chain risk company.
2025 In July，Anthropic secured a contract worth 200000000 USD from the Pentagon.
Anthropic insists that AI must not be used for mass surveillance of Americans, nor may AI autonomously decide to launch lethal attacks.
This year March, the Pentagon placed Anthropic on the blacklist, prohibiting the continued use of its technology in defense projects.
Anthropic said the ban had already cost it tens of billions of USD in business.

AI-assisted interpretation

The following is analysis, separate from reported facts. Verify important claims independently.

Anthropic disputes with the Pentagon over Claude's restrictions on military use.Anthropic believes safety red lines should be retained, while the Pentagon demands the ability to use Claude for all lawful military purposes. The court previously sided with the Pentagon, and Anthropic is now seeking rehearing by all the judges.

Why it matters to readers

The case concerns AI how safety restrictions, military use, and government procurement should be coordinated, and could also affect Anthropic's cooperation on defense projects.

This is not an ordinary product dispute, but a legal dispute over whether AI may refuse certain military uses, and whether the government may therefore restrict a company's participation in defense projects.

Risks and unknowns

The evidence does not indicate whether the U.S. Court of Appeals for the District of Columbia Circuit will accept the en banc rehearing petition.
The evidence provides no court documents or independent sources to verify the claims.
The evidence does not indicate the final ruling.
“tens of billions in USD business” is Anthropic's claim; the evidence provides no specific amount or verification materials.

Anthropic applies for en banc rehearing by the U.S. Court of Appeals over the national security risk designation

Article

AI-assisted interpretation

Why it matters to readers

Risks and unknowns

Related Developments

Related concepts

Keep following

Security Center