Anthropic and Pentagon argue over military use of Claude models

The Pentagon is pushing AI companies including Anthropic to allow the U.S. military to use their technology for all lawful purposes. Anthropic has pushed back more than the others, according to a report in Axios. The government has made the same request to OpenAI, Google, and xAI. An anonymous Trump administration official told Axios that one of those companies has agreed while the other two have shown some flexibility. The Pentagon is reportedly threatening to end its 200 million dollar contract with Anthropic as a result of the resistance. In January the Wall Street Journal reported significant disagreement between Anthropic and Defense Department officials over how its Claude models could be used. The publication later said that Claude was used in the U.S. military operation to capture then Venezuelan president Nicolas Maduro. A company spokesperson told Axios that Anthropic has not discussed the use of Claude for specific operations with the Department of War. The company is instead focused on usage policy questions including its hard limits around fully autonomous weapons and mass domestic surveillance. Anthropic did not immediately respond to a request for comment from TechCrunch.