Anthropic CEO Dario Amodei has confirmed that his AI startup remains committed to its ethical guardrails, even after the Pentagon halted its use of Anthropic’s Claude AI model.

The dispute centers on Anthropic’s refusal to allow its AI to be used for mass surveillance on Americans or to operate fully autonomous weapons without human oversight.

The US Department of Defense demanded that Anthropic permit the use of Claude AI for “all lawful purposes.” When the company refused, citing its two non-negotiable “red lines,” President Trump ordered federal agencies to immediately stop using Anthropic’s technology.

Defense Secretary Pete Hegseth described Anthropic as a “supply chain risk”, instructing military contractors to cease working with the firm. Trump also criticized the company as a “radical left, woke” organization, alleging that it put American lives and national security at risk.

Amodei responded, saying Anthropic’s position is consistent with national security goals and ethical responsibility. He called the government’s actions “retaliatory and punitive” and emphasized that the company has not formally received a supply chain risk designation.

Why Anthropic won’t compromise

Amodei told CBS News:

“We have these two red lines. We are still advocating for those red lines. We’re not going to move on them.”

He argued that AI’s potential for misuse—such as mass surveillance or fully autonomous lethal operations—requires careful oversight. Even though Anthropic is patriotic and wants to assist the military, the company refuses to supply technology that could unintentionally harm civilians or soldiers.

Autonomous weapons, Amodei noted, remain unreliable and lack clear accountability:

“We don’t want to sell something that could get our own people killed or innocent people killed.”

The Pentagon’s Position

Pentagon officials maintain that federal law and internal military policies already restrict mass surveillance and autonomous weapons, so they see no need for written guardrails from Anthropic.

Emil Michael, Pentagon CTO, said:

“At some level, you have to trust your military to do the right thing… But we do have to be prepared for what China is doing.”

The Pentagon reportedly offered written acknowledgements of existing laws, but Anthropic said the legal language was vague enough to potentially bypass the company’s red lines.