Two people familiar with the US military’s use of artificial intelligence confirm that the US employed Anthropic’s Claude AI model for the attack on Iran over the weekend — and continues to use it.
The Pentagon has not specified how the AI tool is being used, but it is being utilized despite a government-wide ban on the technology imposed following a disagreement with the Pentagon last week. The disagreement revolved around Anthropic’s call for safeguards that would explicitly prohibit the military from using Claude to perform mass spying on Americans or to power fully autonomous weapons.
It’s unknown if Claude is being used by the Israeli army in this war as well. A request for comment was not answered by an IDF representative. The IDF has its own “Lavender” targeting system, which it employed in the Gaza conflict, and it does employ AI in combat.
The Pentagon insisted that Claude be allowed to be used for “all lawful purposes” and argued that Anthropic’s usage concerns were irrelevant because the Pentagon is already prohibited from conducting widespread surveillance of Americans due to internal policies that prohibit the use of fully autonomous weapons.
Beyond such regulations, “At some level, you have to trust your military to do the right thing,” stated Emil Michael, the chief technology officer at the Pentagon, in an interview on Friday.
Because “we believe that crossing those lines is contrary to American values, and we wanted to stand up for American values,” Anthropic CEO Dario Amodei told that the company has tried to set “red lines” in the government’s use of its technology.
“Disagreeing with the government is the most American thing in the world,” remarked Amodei. We are also patriots. We have defended the principles of our nation in all that we have done here.”
Defense Secretary Pete Hegseth has identified Anthropic as a supply chain risk, and President Trump said on Friday that he is directing government agencies to cease utilizing the company’s technology, giving them six months to phase it out.
Defense One, citing various sources acquainted with the Department of Defense’s dispute with Anthropic, reported that replacing Claude’s capabilities with another AI platform might take three months or longer.
Michael, the Pentagon’s top technology officer, told that Claude is used by the Defense Department for document synthesis and improving logistics and supply chain efficiency.
