Anthropic Upholds Guardrails in Pentagon AI Standoff

Anthropic CEO Dario Amodei reiterates two guardrails—blocking mass surveillance of Americans and prohibiting fully autonomous weapons—as a condition for any military collaboration, even as the Pentagon pushes for broader, unguarded use of Claude. The dispute followed Trump-era contract cuts and a Defense Department push to phase Anthropic out within six months; Amodei argues the guardrails reflect U.S. values and safety, while Pentagon officials push for more permissive use. The sides have not reconciled, and Congress may weigh in on AI safeguards.