Dario Amodei, CEO of artificial intelligence firm Anthropic, finds himself at the center of a new kind of storm. The company he co-founded has been banned from federal government use, with accusations of it having crossed red lines regarding mass surveillance and fully-autonomous weapons.
Claude, an AI chatbot created by Anthropic for both work and education, was deeply embedded in military intelligence operations within the Pentagon since last summer. Following a recent demand to share its AI capabilities unrestrictedly for lawful military use, Anthropic refused.
“Given all we’ve done for U.S. national security,” Amodei said emphatically, “it feels very punitive and inappropriate.”
Ammodei asserts that their two core principles have remained constant: preventing mass surveillance of Americans and ensuring fully-autonomous weapons do not operate without human oversight.
“These are red lines,” he stated, “and we’ve had them from the start. We still advocate for those red lines.”
For Amodei, who sees himself in a free-market system where different entities offer varying products based on their principles, Anthropic’s capabilities and limitations have been well-understood.
“I think we are a good judge of what our models can do reliably and what they cannot,” he argued. “We believe that the AI should remain under human control.”
His words echo in the current context: the Trump administration has directed all government use of Anthropic’s AI to cease, canceling over $200 million in federal contracts. Defense Secretary Pete Hegseth labeled Anthropic as a “supply chain risk to national security,” marking the first such designation for an American company.
Amodei disputes any notion that Anthropic operates with a bias or partisanship, insisting instead on its adherence to neutrality and fair principles. Criticized by some for seeming partisan, Amodei reiterates their commitment to unbiased support of national security measures.
“Everything we have done has been for the sake of this country,” he reiterated. “We believe in defeating our autocratic adversaries and defending America.”
Ammodei stands firm, planning to pursue legal action against the ban while maintaining a stance at the negotiation table. His main rival, Sam Altman from OpenAI, reached an agreement with Pentagon officials on Friday.
As Amodei faces a government ban, critics argue it represents an abuse of power. “This has never happened before,” he asserts. “We have tried to be very neutral and stand up for American values.”
The Trump administration’s decision highlights the ongoing debate over who should control advanced technology: private companies or the federal government.