Arabic version: أنثروبيك ترفض مطالب البنتاغون بشأن ضمانات استخدام الذكاء الاصطناعي

Anthropic has firmly stated that it will not comply with the U.S. Department of Defense’s (DoD) requests regarding the usage of its artificial intelligence (AI) technology. Chief Executive Dario Amodei emphasized that the company prefers to sever ties with the Pentagon rather than compromise on what it sees as essential democratic values.

According to BBC News, Amodei’s comments followed a meeting with U.S. Secretary of Defense Pete Hegseth, where the DoD demanded that Anthropic allow “any lawful use” of its AI tools. The discussion concluded with a warning that the Pentagon might remove Anthropic from its supply chain if it did not comply.

The core issues for Anthropic revolve around the potential applications of its AI systems, particularly in mass domestic surveillance and fully autonomous weapons. Amodei stated that these uses have never been part of their agreements with the Department of War and should not be included now.

Despite receiving updated contract language from the DoD, an Anthropic spokeswoman noted that it offered little progress in preventing the misuse of its technology. She criticized the new wording as a superficial compromise that could allow safeguards to be ignored.

The disagreement has escalated, with Emil Michael, the U.S. Undersecretary for Defense, publicly criticizing Amodei, accusing him of jeopardizing national security. The Pentagon has indicated that if Anthropic does not comply, it may invoke the Defense Production Act, which allows the government to prioritize defense needs.

Amodei reiterated that while Anthropic supports lawful foreign intelligence missions, using AI for mass surveillance contradicts democratic principles. He also expressed concerns about the reliability of AI systems for military applications, stating that they are not dependable enough for fully autonomous weapons without proper oversight.