The announcement came hours after US President Donald Trump ordered federal agencies to halt the use of AI systems developed by Anthropic, citing concerns over military deployment and the company’s attempts to impose conditions on the Pentagon.
Altman highlighted that the Pentagon had “displayed a deep respect for safety and a desire to partner to achieve the best possible outcome.” He added that two core safety principles – prohibitions on domestic mass surveillance and human responsibility for the use of force, including autonomous weapon systems – are reflected in the agreement. The company will also deploy technical safeguards and full-disk encryption to ensure AI models function securely, but only on cloud networks, he added.
The OpenAI CEO called on the Pentagon to extend these terms to all AI companies, urging Washington to move away from taking legal action.
Previously, the Pentagon had relied on Anthropic’s Claude model in classified networks. However, Anthropic later refused to grant full access to its AI systems without guarantees that the technology would not be used for domestic surveillance or fully autonomous weapons targeting without human oversight.
Pentagon officials saw this as an attempt by the company to dictate terms beyond a normal business partnership. Deputy Defense Secretary Emil Michael criticized Anthropic CEO Dario Amodei for trying to exert undue control over military operations, escalating tensions Bloomberg reported earlier this week.
After banning government agencies from using Anthropic’s AI developments for six months, Trump accused the company of trying to pressure the Pentagon into complying with its conditions and of putting national security at risk.
“We don’t need them, we don’t want them, and we will no longer deal with them,” he wrote on Truth Social.