Anthropic says it will not allow its AI to be used without limits by the US military, citing risks including mass surveillance and fully autonomous weapons. CEO Dario Amodei said using these systems for “mass domestic surveillance is incompatible with democratic values,” and argued the technology is not yet reliable enough to be trusted to power lethal systems without a human ultimately in control. He added, “We will not knowingly provide a product that puts America’s warfighters and civilians at risk.”
Anthropic was one of several US AI firms—along with OpenAI, Google and Elon Musk’s Grok—that the Pentagon contracted in 2025 to supply models for a range of military applications under an agreement worth $200 million (€170 million). According to reporting, after a meeting with Amodei, Defense Secretary Pete Hegseth pressed Anthropic to open its technology for use in a classified setting by the end of the week or face losing the government contract; the other companies have already done so.
Officials warned they could escalate by labeling Anthropic a supply-chain risk or by invoking the Defense Production Act to broaden government authority over its products. Amodei described those threats as “inherently contradictory,” saying they simultaneously brand Anthropic’s systems a security threat and claim they are essential to national security. He also criticized the administration’s posture, saying Anthropic understands that the “Department of War, not private companies, makes military decisions,” and warned that AI can sometimes undermine rather than protect democratic values.
A Pentagon spokesman, Sean Parnell, said the military “has no interest in using AI to conduct mass surveillance of Americans (which is illegal) nor do we want to use AI to develop autonomous weapons that operate without human involvement.” Officials confirmed one exchange between Anthropic and the Pentagon touched on intercontinental ballistic missiles, underscoring the sensitivity of the dispute. Parnell argued that opening full access would be necessary to avoid “jeopardizing critical military operations” and said the department “will not let ANY company dictate the terms regarding how we make operational decisions.”
The standoff has drawn political attention. Senator Mark Warner, the top Democrat on the Senate Intelligence Committee, said he was “deeply disturbed” by reports that the Pentagon is “working to bully a leading US company.” Warner said the episode illustrates that the Department of Defense is not adequately addressing AI governance and called for Congress to pass strong, binding AI rules for national security contexts.
The dispute highlights a broader tension over how commercial AI systems should be governed when they intersect with national security: private-sector safety and democratic norms versus military requirements and operational control. Anthropic’s decision signals the company’s willingness to forgo certain defense partnerships rather than cede limits it views as essential to safety and civil liberties.