美国哥伦比亚特区上诉法院以2:1的判决维持了国防部对Anthropic公司的供应链风险认定，驳回了这家AI公司关于该认定武断、越权和违宪的主张 1。法院认定国防部有充分理由认为Claude模型继续整合进国防部信息系统构成国家安全风险 1。
国防部在2025年3月将Anthropic列为供应链风险，禁止美国军方和国防承包商使用其Claude模型 1。这一认定源于双方在军事合作中的分歧。Anthropic与国防部曾于2025年7月签署2.5亿美元合同，但在2025年9月谈判破裂 1。分歧的焦点在于，国防部要求不受限制地访问Claude模型，而Anthropic则坚持不将该技术用于全自主武器或国内大规模监控 1。国防部长Pete Hegseth指控Anthropic试图"夺取美国军方运作决策的否决权" 1。
法官Gregory Katsas在多数意见中表述，国防部有充分理由认为Claude继续整合进国防部信息系统构成法定涵盖的国家安全风险 1。Anthropic保留进一步上诉的权利，包括向最高法院申请 1。
A U.S. appeals court in Washington, D.C. issued a 2-to-1 ruling Friday upholding the Department of Defense's classification of Anthropic as a supply chain risk, rejecting the AI company's claims that the Pentagon's ban on its Claude model was arbitrary, exceeded authority, and violated the Constitution 1. The DoD designated Anthropic a supply chain threat in March 2025, prohibiting U.S. military and defense contractors from using its models after negotiations between the two parties collapsed 1.
The conflict emerged from disagreement over the terms of collaboration. The Pentagon and Anthropic had signed a $250 million contract in July 2025 but broke apart in September 2025 over restrictions on model access 1. The Department of Defense sought unrestricted access to Claude, while Anthropic insisted the technology not be deployed for fully autonomous weapons or domestic mass surveillance 1. Defense Secretary Pete Hegseth accused Anthropic of attempting to seize "veto power over U.S. military operational decisions" 1.
The appellate judges determined that the Pentagon had sufficient grounds to view continued integration of Claude into DoD information systems as a legally cognizable national security risk 1. Anthropic retains the option to pursue further appeals, including a petition to the Supreme Court 1. The decision follows a separate ruling last month from a federal judge in San Francisco that struck down another parallel Pentagon supply chain risk designation 1.
评论
还没有评论，欢迎留下第一条。