The Pentagon is close to cutting ties with the artificial intelligence company Anthropic. Reports say the move comes from frustration over limits on using Anthropic’s Claude AI tool. Talks to renew the contract are stalled because Anthropic wants extra safeguards. These safeguards are meant to stop Claude from being used for mass surveillance of Americans or creating weapons that operate without humans. A source familiar with the talks, who asked to stay anonymous, said the Pentagon wants to use Claude as long as the law is not broken. Earlier, Axios reported on this disagreement. The use of AI in weapons and personal data gathering is a growing concern. Anthropic positions itself as a responsible AI company that avoids catastrophic harm from its tech. It built a special Claude Gov version for US national security with enhanced skills for handling classified info and cybersecurity data. Still, the Pentagon sees Anthropic’s restrictions as a possible supply chain risk and may label the company accordingly.