Anthropic Refuses Pentagon Demand Over AI Use
Anthropic's CEO said the company won't permit Claude's use for mass domestic surveillance or fully autonomous weapons as a Pentagon Friday 5:01 p.m. deadline looms.
Overview
Anthropic CEO Dario Amodei said Thursday the company "cannot in good conscience accede" to Pentagon demands and would rather not work with the Defense Department than allow certain uses of its AI.
The Pentagon has demanded access to Anthropic's Claude model for "all lawful purposes" and set a Friday 5:01 p.m. deadline, warning it will terminate the partnership or deem the company a supply chain risk.
Pentagon spokesman Sean Parnell said the department will not let any company dictate how it makes operational decisions.
Pentagon chief technology officer Emil Michael said federal law and Pentagon policies already bar mass surveillance of Americans and fully autonomous weapons and invited Anthropic to participate in its AI ethics board.
Anthropic was awarded a $200 million contract last summer and deployed systems on classified networks via Palantir, and Pentagon officials said they could invoke the Defense Production Act if negotiations fail.
Analysis
Center-leaning sources frame the story as a principled corporate stand against military overreach by emphasizing Anthropic's ethical language (e.g., 'democratic values') and using loaded terms like 'demand' and 'Department of War.' Editorial choices privilege company claims and a critical anonymous ex-DoD voice while the Defense Department's perspective is notably absent; quoted material remains source content.
Sources (18)
FAQ
Anthropic seeks explicit prohibitions against using Claude for mass surveillance of Americans or for fully autonomous weapons without human involvement.
The Pentagon set a Friday 5:01 p.m. deadline, threatening to terminate the partnership, designate Anthropic a supply chain risk, and potentially invoke the Defense Production Act.
Anthropic was awarded a $200 million contract last summer to prototype AI capabilities for national security, with Claude deployed on classified networks via Palantir.
OpenAI, Google, and xAI are working collaboratively with the Pentagon for 'all lawful purposes'; xAI's Grok is approved for classified use, and OpenAI and Google are close.
Pentagon officials state that mass surveillance and fully autonomous weapons are already barred by federal law and policies, and emphasize trusting the military while rejecting company-dictated operational rules.
History
This story does not have any previous versions.