Sipu Jha
Anthropic CEO Dario Amodei has issued a firm rejection of Pentagon pressure to loosen safeguards on its frontier AI model, Claude. In a strongly worded blog post, Amodei said there are two demands his company will not accept ‘under any circumstances’ even if it means losing government contracts. “We cannot in good conscience accede to their request,” Amodei wrote. He also warned that the demand made by Pentagon for ‘any lawful use’ would force the company to cross two red lines which are mass domestic surveillance and fully autonomous weapons. For those unaware, Defense Secretary Pete Hegseth gave Anthropic until Friday (February 27) evening to agree to the military terms or risk being backlisted. The officials have also threatened to invoke the Defense Production Act which is a wartime law that would give the president control over Anthropic’s resources and ladle the company a supply chain risk.
The two red lines
- Most dramatic surveillance: Amide believe that using AI for mass surveillance of US citizens is incompatible with democratic values. He also warned that powerful AI systems could automatically assemble scattered data like such as movements, browsing histories, and associations—into detailed profiles at massive scale. “AI-driven mass surveillance presents serious, novel risks to our fundamental liberties,” he wrote, noting bipartisan opposition in Congress to current surveillance practices.
- Fully autonomous weapons: While partially autonomous weapons are already in use , Amodei also mentioned that today’s AI systems are not reliable enough to power fully autonomous weapons that remove humans from the decision loop. “We will not knowingly provide a product that puts America’s warfighters and civilians at risk,” he stated, stressing that such systems lack the oversight and judgment exercised by trained military personnel.
Anthropic CEO noted Pentagon’s contradictory stance
Amodei noted that Pentagon’s stance was contradictory: one threat labels Anthropic a national security risk, while another insists Claude is essential to national defense. He also stressed on the fact that Anthropic has already offered significant value to US agencies by deploying Claude for intelligence analysis, cyber operations and operational planning. He revealed that the company has also has also cut off revenue streams linked to Chinese military firms and supported export controls to protect democratic advantages.
Pentagon’s response
Hours before Amodei’s statement, a Pentagon spokesperson insisted the department had no interest in using AI for mass surveillance of U.S. citizens or for autonomous weapons. However, sources told Business Insider that new contract language introduced this week effectively gave the military discretion to override Anthropic’s safeguards.
English 
हिन्दी