Newly unsealed court documents have offered a rare look inside a souring relationship between Anthropic, one of America's leading artificial-intelligence companies, and the U.S. defense establishment.
According to the Times of India, the documents were unsealed this week and contain emails exchanged between Anthropic chief executive Dario Amodei and the Pentagon's IT head. The correspondence, the report says, illustrates how the AI firm fell out with what the outlet refers to as the Department of War. At some point in that dispute, Anthropic was branded a "national security threat."
The Times of India notes that the emails were first reported by The Wall Street Journal. Beyond the existence of the correspondence, the involvement of Amodei personally, and the harsh "national security threat" characterization, the available reporting does not spell out the specific policy disagreements, timeline, or what either side was ultimately seeking.
What is clear is that a direct channel between a top AI executive and a senior Pentagon technology official broke down badly enough to surface in litigation and unsealed court filings.
Why it matters: AI companies and the U.S. military are increasingly entangled, and how a firm like Anthropic navigates — or clashes with — the Pentagon signals how tightly the government intends to control the powerful technology it also wants to buy.