AI Safety Clash Deepens as Pentagon Pressures Anthropic

Pentagon has given Anthropic an ultimatum to relax safeguards on its Claude AI model or face potential penalties, intensifying debate over ethics and military use of artificial intelligence.

The U.S. Department of Defense has intensified discussions with Anthropic over how its Claude AI model can be used in military settings.

Defense Secretary Pete Hegseth met with Anthropic CEO Dario Amodei to address restrictions the company has placed on certain applications of Claude. The discussions reportedly focused on limits that prevent the model from being used in autonomous weapons systems and certain surveillance-related contexts.

U.S. officials have pressed Anthropic to reconsider those safeguards. Reports indicate that the Pentagon has set a deadline for the company to respond and has discussed potential consequences if the restrictions remain unchanged.

The U.S. government has previously worked with Anthropic under defense-related arrangements involving AI systems. Claude has been deployed in certain classified environments, though the company has maintained usage guardrails aligned with its safety policies.

Also read: After OpenAI, Anthropic Flags Chinese AI Firms Over Distillation

The dispute comes as the Defense Department continues expanding its use of artificial intelligence tools in national security operations. It also follows broader scrutiny of advanced AI companies regarding how their systems are accessed and deployed in military contexts.

Anthropic has consistently emphasized safety and responsible AI deployment. As of now, the company has not announced changes to its published usage policies.

The situation reflects growing friction between AI developers that maintain operational safeguards and government agencies seeking broader access to advanced AI systems for defense applications.

Do not miss:
Founder Warns of AI Shift in Marketing After Claude “Killed” Her Startup
AI Godfather Geoffrey Hinton Warns AI Will Replace Millions of Jobs