CEO Dario Amodei says humanity “cannot in good conscience consent” to the military use of AI.

Washington — Anthropic CEO Dario Amodei said Thursday that the artificial intelligence company “cannot in good conscience comply” with the Pentagon’s request to allow broader use of its technology.

In a statement, the company said it was not withdrawing from negotiations, but said the new contract language it received from the Pentagon “made virtually no progress in preventing mass surveillance of Americans or its use in fully autonomous weapons.”

The Pentagon’s top spokesperson reiterated that the military wants to use Anthropic’s artificial intelligence technology in a lawful way and does not intend to impose any restrictions on the company ahead of Friday’s deadline for consent.

Sean Parnell said on social media Thursday that the Pentagon “has no interest in using AI for mass surveillance of US citizens, nor does it intend to use AI to develop autonomous weapons that operate without human involvement.”

Anthropic’s policies prohibit models such as the chatbot Claude from being used for those purposes. It is the last of its kind, with the Pentagon also having deals with Google, OpenAI and Elon Musk’s xAI, but not providing its technology to new internal US military networks.

Read more: Reddit sues AI company Anthropic for allegedly ‘scraping’ user comments to train chatbot Claude

Parnell said the Pentagon wants to “use Anthropic models for any lawful purpose,” but did not provide further details. He said opening up the use of the technology would prevent the company from “putting critical military operations at risk.”

“We will not allow any company to dictate the terms of how we make business decisions,” he said.

At Tuesday’s meeting between Defense Secretary Pete Hegseth and Anthropic CEO Dario Amodei, military officials warned that they could designate Anthropic as a supply chain risk, terminate the contract or, even if the company doesn’t approve, invoke a Cold War-era law called the Defense Production Act to give the military broader authority to use the company’s products.

Parnell mentioned just two of those impacts in Thursday’s post about X, saying Anthropic “will make a decision by 5:01 p.m. ET on Friday.”

“Otherwise, we will terminate our partnership with Anthropic and consider it a supply chain risk,” he wrote.

Anthropic did not immediately respond to a request for comment Thursday. “Consistent with what our model can do reliably and responsibly, we have continued to discuss our usage policies in good faith to ensure that Anthropic continues to support the government’s national security mission,” the company said in a statement after Tuesday’s meeting.

Read more: Facebook owner Meta buys AI chips from AMD in deal worth up to $100 billion

Sen. Thom Tillis, R-North Carolina, who is not seeking re-election, said Thursday that Anthropic is “doing everything in its power to help us from ourselves,” while the Pentagon is handling the issue unprofessionally.

“Why on earth are we having this discussion in public?” Tillis told reporters. “This is not how you should deal with strategic vendors that you have contracts with.”

He added: “When companies are resisting market opportunities out of fear of negative repercussions, it’s time to listen to them and find out what they’re really trying to solve behind closed doors.”

Sen. Mark Warner (Virginia), the ranking Democrat on the Senate Intelligence Committee, said he was “deeply disturbed” by reports that the Pentagon is “trying to bully some of America’s leading companies.”

“Unfortunately, this is further evidence of the Department of Defense’s complete disregard for AI governance,” Warner said in a statement. This “further highlights the need for Congress to enact strong and binding AI governance mechanisms from a national security perspective.”

Although Pentagon officials have said they always follow the law when it comes to using AI models, Hegseth told Fox News last February, just weeks after taking over as defense secretary, that “Ultimately, we want lawyers who are going to give us good constitutional advice and not be there to get in the way of anything.”