Is Anthropic a 'supply-chain risk'? US judge says otherwise
What's the story
A US district judge said during a hearing yesterday that the Donald Trump administration has not provided sufficient evidence to classify AI firm Anthropic as a 'supply-chain risk.' The comments came amid stalled contract negotiations between Anthropic and the Department of Defense (DoD). The judge's remarks were first reported by Bloomberg and Axios.
Concerns
Concerns over AI's use in mass surveillance, lethal weapon targeting
Anthropic has raised concerns over its AI being used for mass surveillance of Americans or lethal weapon targeting and firing decisions.
The company argues that the technology isn't ready for such applications.
However, the Pentagon has countered these claims, saying a private company shouldn't dictate how military technologies are used and promising to use them in "lawful" ways.
Retaliation
Judge warns against potential precedent for retaliating
The US government also contended that Anthropic's public criticism of the DoD justified the ban.
However, US District Judge Rita Lin found this reasoning "really troubling," warning it could set a precedent for retaliating against federal contractors who disagree with the administration.
The judge had temporarily blocked the ban in March and is now considering whether to make that order permanent.
AI control
No evidence that AI can be altered or disabled: Experts
The DoD also claimed that Anthropic could disable or alter its AI models during warfighting operations.
However, experts have said there is no evidence to support this claim.
Judge Lin agreed with the experts, saying she saw no proof that Anthropic could alter a delivered model or "flip some kind of kill switch."