
Pentagon may have used OpenAI models via Microsoft Azure before military ban lift
India Today
OpenAI reached an agreement to deploy its AI systems for the US Department of Defense a few days ago. However, a new report reveals that the Pentagon may have used OpenAI's models as early as 2023 via Microsoft.
OpenAI has found itself in a tough spot after reaching an agreement with the Pentagon to deploy its AI systems. The agreement, which OpenAI CEO Sam Altman later admitted was rushed, was met with widespread backlash online, with many ditching ChatGPT as a way to express their disapproval. Now, a new report indicates that the Pentagon may have had access to OpenAI’s models as early as 2023 via Microsoft’s Azure.
According to Wired, in 2023, when OpenAI's usage policy explicitly banned military applications, the Pentagon began experimenting with OpenAI's models via Microsoft's Azure cloud platform.
The report states that the US Department of Defense accessed Azure OpenAI before OpenAI lifted its embargo on military use. At the time, Microsoft held a broad license to commercialise OpenAI's technologies. The Redmond giant has had longstanding contracts with the Department of Defense for years now.
According to the report, OpenAI employees spotted Pentagon officials in the company's San Francisco offices during this period. There was confusion internally regarding the scope of OpenAI's policy. It was unclear if Microsoft's Azure OpenAI products fell under OpenAI's restrictions.
Microsoft spokesperson, Frank Shaw, told Wired, "Microsoft has a product called the Azure OpenAI Service that became available to the US Government in 2023 and is subject to Microsoft terms of service."
The company did not comment on whether the AI platform was used by the Pentagon specifically. However, Microsoft claims that the service only got approved for “top secret” government work in 2025.













