US Court Upholds Pentagon's Security Risk Label on Anthropic
A federal appeals court in Washington, D.C. ruled that the US Department of Defense lawfully classified AI company Anthropic as a 'supply chain risk.' Judges sided with the government in a 2-1 decision.
The dispute began when Anthropic refused to give the Pentagon unrestricted access to its AI model, Claude. The company wanted to contractually exclude uses like mass surveillance of US citizens and fully autonomous weapons systems.
The Pentagon rejected these conditions and ordered federal agencies to stop using Anthropic's technology. According to court documents, the department began removing the AI systems in March.
The judges wrote, according to Ars Technica, that it was 'deeply sobering' that overly constrained AI models could shut down unexpectedly, causing important military operations to fail. Anthropic had countered that unconstrained models might wrongly identify targets for lethal force.
An Anthropic spokesperson told CNBC the company disagrees with the ruling and is considering all options, including further legal review. A separate federal court in California had ruled a similar designation unlawful in August.
Because the government used two different legal grounds for the designation, two courts had to rule separately. What happens next is unclear: Anthropic can request a new hearing or take the case to the Supreme Court.
WHY IT MATTERS
The case shows how governments can try to pressure AI companies into giving up ethical limits on surveillance and autonomous weapons use.
WHAT YOU CAN DO NOW
- Stay informed on AI policy news Follow reliable news sources covering government decisions about AI use. This helps you notice early when rules on surveillance or weapons systems change.
- Question how agencies use AI When public agencies use AI systems, look into what exactly they're used for and what limits apply. Civil rights groups often publish details on this.
- Support calls for AI transparency Advocate for clear laws defining how AI can be used in military or surveillance contexts. You can do this through petitions or civic engagement.
- Check privacy terms of AI tools you use Read the terms of service for AI apps you use, especially regarding data sharing with authorities. This tells you what protections apply to your data.