Pentagon Tested OpenAI Models Before Military Ban Was Lifted
OpenAI employees allege the DOD used Microsoft Azure to access OpenAI tech while the company still banned military use.
The Pentagon apparently didn't wait for permission. Multiple OpenAI employees allege that the Department of Defense experimented with OpenAI's models through Microsoft's Azure platform before OpenAI officially dropped its blanket ban on military applications in January 2024.
The timeline matters here. OpenAI had an explicit prohibition on military use of its technology. Microsoft, however, offered OpenAI's models through its Azure cloud service — creating a convenient side door for defense applications.
OpenAI lifted that military use ban in January 2024, opening the door to Pentagon partnerships. But according to sources cited by Wired, the DOD was already kicking the tires well before the policy changed.
The revelations raise uncomfortable questions about the gap between OpenAI's stated policies and how its technology was actually being accessed through Microsoft's infrastructure. The Azure loophole, it seems, was doing heavy lifting.