WASHINGTON (AP) – President Donald Trump announced Friday that he is ordering all federal agencies to stop using Anthropic’s technology in response to the company’s unusual circumstances. Public dispute with the Department of Defense Concerning the safety of artificial intelligence.
Trump’s comments came just over an hour before a Pentagon deadline for Anthropic to allow unrestricted military use of its artificial intelligence technology or face consequences — and nearly 24 hours after CEO Dario Amodei said his company. “I cannot in good conscience consent.” In response to the request of the Department of Defense.
President Trump said most government agencies must immediately stop using Anthropic’s AI, but gave the Pentagon six months to phase out the technology already integrated into military platforms.
“We don’t need it, we don’t want it, and we’ll never do business with them again!” Trump wrote.
At issue in the defense contract were conflicts over the role of AI in national security and concerns about how increasingly sophisticated machines would be used in high-stakes situations such as lethal force, classified intelligence, and government surveillance.
The move could benefit Elon Musk’s rival chatbot Grok, which the Pentagon plans to give access to classified military networks, and could also serve as a warning to two other competitors, Google and OpenAI, which also have contracts to supply AI tools to the military.
AP AUDIO: Anthropic refuses to bow to Pentagon on AI safeguards as litigation deadline approaches
Associated Press correspondent Ben Thomas reported that AI companies are rejecting the Pentagon’s request.
Anthropic, the maker of the chatbot Claude, could have lost its contract. But the ultimatum issued by Secretary of Defense Pete Hegseth this week poses broader risks at the height of the company’s meteoric rise from a little-known computer science lab in San Francisco to one of the world’s most valuable startups.
Humans rejected the Pentagon’s final proposal on security measures
If Amodei did not relent, military officials said they would not only terminate Anthropic’s contract but also “deem it a supply chain risk.” This designation is typically imprinted on foreign adversaries and can derail important partnerships between the company and other companies.
President Trump made no such designation in Friday’s announcement, but said Anthropic could face “significant civil and criminal consequences” if it is not served during the phase-out period. Antropic did not immediately respond to a request for comment on Trump’s remarks.
And if Mr. Amodei had relented, he might have lost the trust of the burgeoning AI industry, especially the top talent drawn to the company by its promise to responsibly develop better-than-human AI that could pose catastrophic risks without safeguards.
Anthropic said it has asked the Pentagon for limited assurances that Claude will not be used for mass surveillance of Americans or for fully autonomous weapons. But after months of private discussions that escalated into public debate, the company said in a statement Thursday that the new contract language was “framed as a compromise, combining these safeguards with legal language that can be freely ignored.”
It came after top Pentagon spokesperson Sean Parnell posted on social media that the military “has no interest in using AI to conduct mass surveillance of Americans (which is illegal), nor does it want to use AI to develop autonomous weapons that operate without human involvement.”
He emphasized that the Department of Defense “wants to use Anthropic’s models for the following purposes.” all lawful purposes”, but did not elaborate on how he or other officials hope to use the technology.
Conflict further polarizes the tech industry
In Silicon Valley, a growing number of technology officials from Anthropic’s biggest rivals, OpenAI and Google, expressed support for Amodei’s position in an open letter late Thursday.
OpenAI and Google, along with Musk’s xAI, have a deal to supply their AI models to the military.
Musk on Friday endorsed the Trump Republican administration, saying on his social media platform X that “humanity hates Western Civilization.” All major AI models, including Musk’s Grok and OpenAI’s ChatGPT, are programmed with a set of instructions that guide the chatbot’s values and behavior. In anthropology, this guideline is called the constitution.
Some technology leaders allied with Trump, including Musk and Palmer Lackey, co-founder of defense contractor Anduril, have joined the fray, but the polarizing debate over “woke AI” has also put other companies in a difficult position.
“The Department of Defense is negotiating with Google and OpenAI to agree to what Anthropic has rejected,” reads an open letter from some OpenAI and Google employees. “They’re trying to break up each company because they’re afraid the other will give in.”
But in a surprise move by one of Amodei’s most powerful rivals, OpenAI CEO Sam Altman sided with Anthropic in an interview on CNBC on Friday, questioning the Pentagon’s “threatening” move and suggesting that OpenAI and much of the AI field share the same red line. Mr. Amodei previously worked at OpenAI until leaving in 2021 to found Anthropic with other OpenAI leaders.
“Despite our differences with Anthropic, I pretty much trust them as a company and I think they really care about safety,” Altman told CNBC. “I’m glad they supported our warfighters. We don’t know how this situation will turn out.”
Some lawmakers and former Pentagon AI leaders push back
Some members of Congress and the former leader of the Pentagon’s AI initiative also expressed concerns about the Pentagon’s approach.
Sen. Mark Warner (Virginia), the top Democrat on the Senate Intelligence Committee, said the move “combined with the inflammatory rhetoric attacking the company raises serious concerns about whether national security decisions are made with careful analysis or political considerations.”
Retired Air Force Gen. Jack Shanahan wrote on social media that “hitting the mark on humanity’s problems attracts vitriolic headlines, but in the end everyone loses.”
Shanahan said Claude is already widely used across the government, including in classified settings, and Anthropic’s red line is “reasonable.” He said the large-scale language models of AI that power chatbots like Claude are also “not yet ready for prime time in national security settings,” especially for fully autonomous weapons.
“They’re not trying to be cute here,” he wrote on LinkedIn on Thursday.
When Hegseth and Amodei met on Tuesday, military officials warned they could be named. human Cancel contracts or invoke Cold War-era laws as supply chain risks defense production act The idea is to give the military broader authority to use its products even if the company doesn’t approve them.
“The latter two threats are inherently contradictory. One sees us as a security risk, while the other sees Claude as essential to national security,” Amodei said Thursday. He said he hopes the Pentagon will reconsider given Claude’s value to the military, but if not, Anthropic will “work to enable a smooth transition to another provider.”
___
O’Brien reported from Providence, Rhode Island.
