OpenAI Secures Pentagon Deal Amid AI Safety Concerns

SAN FRANCISCO — In a significant development, OpenAI’s CEO, Sam Altman, revealed late Friday that the company has entered into an agreement with the Pentagon for the utilization of its AI technologies within the military’s classified systems. This deal appears to align with similar safety protocols that rival company Anthropic has been advocating for.

The announcement coincides with President Donald Trump’s directive requiring all federal agencies to discontinue the use of Anthropic’s AI tools. Defense Secretary Pete Hegseth characterized Anthropic as a “supply chain risk” due to its unwillingness to comply with the Pentagon’s demands regarding limitations on its AI systems, particularly concerning autonomous weaponry and mass surveillance of U.S. citizens.

Altman’s remarks indicate that the Pentagon has agreed to impose similar restrictions on OpenAI’s models. He stated, “Two of our most important safety principles are prohibitions on domestic mass surveillance and ensuring human accountability for the use of force, including in relation to autonomous weapon systems. The Department of War (DoW) endorses these principles, integrates them into law and policy, and we have included them in our agreement.” He further detailed that technical safeguards will be installed to ensure the models operate as intended, which is a requirement from the DoW.

Moreover, Altman mentioned plans to deploy engineers to the Pentagon to oversee the security of their models. He also expressed a desire for the DoW to extend these same terms to other AI firms, emphasizing, “We believe all companies should be amenable to these conditions. Our preference is for a de-escalation from legal confrontations and towards reasonable agreements.”

In response to the situation, Anthropic announced its intentions to legally contest the “supply chain risk” label, a designation usually applied to businesses with affiliations to foreign adversaries. As a result, all contractors collaborating with the military would be required to demonstrate that their military engagements do not involve Anthropic’s products.

Currently, the specifics of OpenAI’s arrangement with the Pentagon compared to Anthropic’s requests remain unclear.

CNN has reached out to both the Pentagon and OpenAI for further clarification.

Hegseth shared Altman’s announcement on X, and Under Secretary Emil Michael, who oversees technology at the Pentagon, added in a separate X post, “In matters of life and death for our warfighters, having a reliable partner that engages in good faith is crucial as we move into the AI Age.”

The-CNN-Wire & 2026 Cable News Network, Inc., a Warner Bros. Discovery Company. All rights reserved.