The Pentagon is pressuring AI development companies to grant permissions to use their technologies by the U.S. military for all lawful purposes, but Anthropic, according to a new Axios report, is resisting.
According to reports, the government is making the same demands of OpenAI, Google, and xAI. An anonymous official from the Trump administration told Axios that one of these companies agreed, while two others, according to the source, showed some flexibility.
Anthropic, meanwhile, is reportedly the least willing to concede. In response, the Pentagon allegedly threatens to halt its $200 million contract with the AI company.
In January The Wall Street Journal reported significant disagreements between Anthropic and Department of Defense officials about how Claude models could be used. Subsequently, data emerged about the use of Claude in U.S. operations against Venezuela during the arrest of Nicolás Maduro.
Negotiations on Claude usage between Anthropic and the government
“they did not discuss using Claude for specific operations with the Department of War”
– Axios
A spokesperson for Axios noted that the company “did not discuss using Claude for specific operations with the Department of War,” but focused on a specific set of policy-use questions – in particular on our strict limits on fully autonomous weapons and mass surveillance.
“focused on a specific set of policy-use questions – specifically on our strict limits on fully autonomous weapons and mass surveillance”
– Axios
At the same time, analysts point out that the government continues to discuss further steps regarding Claude’s use in military applications and the responses of the developers to the requirements for deploying their systems in defense and national security.