Anthropic Faces Pentagon Pressure to Grant Military Access to AI Model
In a escalating dispute, Anthropic has until Friday evening to decide whether to provide the U.S. military unrestricted access to its advanced AI model, according to Axios. Failure to comply may result in significant repercussions as Defense Secretary Pete Hegseth has indicated that the Pentagon could classify Anthropic as a “supply chain risk”—a designation typically reserved for foreign adversaries. Alternatively, the government may invoke the Defense Production Act (DPA) to compel the company to create a version of its AI tailored to military requirements.
The DPA empowers the President to mandate companies to prioritize production for national defense, a power `recently utilized during the COVID-19 pandemic to expedite the manufacture of essential medical supplies. Anthropic has consistently expressed its opposition to the use of its technology for mass surveillance or fully autonomous weapons, holding firm on these core principles.
Pentagon officials assert that military technology usage should adhere to U.S. laws and constitutional limits rather than the policies of private contractors. Experts are concerned that using the DPA in this context would represent a notable expansion of its application, showcasing a broader trend of increasing instability within the executive branch. Dean Ball, a senior fellow at the Foundation for American Innovation, has remarked that this could set a concerning precedent, making it appear as if the government might retaliate against companies that oppose its political stance.
The situation reflects deeper ideological divides within the administration, with some officials, including AI czar David Sacks, critiquing Anthropic’s safety policies as overly cautious. Business leaders are increasingly alarmed; as Ball noted, this instability could undermine America’s reputation as a reliable environment for global commerce, which relies heavily on a stable and predictable legal framework.
Meanwhile, Anthropic appears resolute, with sources indicating that the company has no intention of relaxing its usage restrictions. Notably, it remains the only frontier AI lab with classified access to the Department of Defense. The Pentagon currently lacks alternative options, having reportedly entered into an arrangement with xAI’s Grok for classified systems, further elucidating the aggressive stance being taken.
This critical conflict underscores the challenges faced by the Pentagon in securing reliable partnerships in the evolving landscape of artificial intelligence, suggesting a high-stakes game of negotiation where Anthropic may choose to stand its ground.
