The Pentagon has issued a Friday 5:01 p.m. deadline to Anthropic regarding unrestricted access to its AI systems, with Defense Secretary Pete Hegseth demanding the company either comply or face consequences.
The standoff represents a significant conflict over the integration of commercial artificial intelligence into national defense operations. The Department of Defense views Anthropic’s technology as essential to national security, while the company maintains that unrestricted access poses risks to democratic values. The outcome could affect ongoing military planning and operations currently utilizing the AI models.
Anthropic CEO Dario Amodei stated he “cannot in good conscience accede to [the Pentagon’s] request.” Amodei identified two specific uses he considers unacceptable: mass surveillance of Americans and fully autonomous weapons with no human in the loop. He argued that some uses are outside the bounds of what current technology can safely and reliably do.
The Department of Defense has threatened to label Anthropic a supply chain risk or invoke the Defense Production Act to compel compliance. Amodei noted a contradiction in these threats, stating one labels Anthropic a security risk while the other labels its AI essential to national security. He emphasized that while the Department makes military decisions, certain AI applications could undermine democratic values.
Amodei expressed a preference to continue serving the Department of Defense with the two requested safeguards in place. He stated that if the Pentagon chooses to offboard Anthropic, the company will work to enable a smooth transition to another provider. The Department is reportedly preparing xAI for similar work.
Anthropic is an AI research company known for its Claude models. The Defense Production Act gives the president authority to force companies to prioritize or expand production for national defense. The company currently provides classified-ready systems to the military.
