Pentagon quietly began testing OpenAI models in 2023 despite military ban: Report
Last week, OpenAI signed an agreement with the US government that allows its AI models to be used by the military for covert use cases. However, a new report by WIRED states that the US military began testing OpenAI models in 2023, despite the company imposing a blanket ban on the military’s access to artificial intelligence models.
Reportedly, OpenAI employees discovered in 2023 that the Pentagon had begun testing its models through Azure OpenAI, a version of OpenAI models offered by Microsoft. The report notes that Microsoft had been under contract with the Department of Defense for decades at the time.
Microsoft, in particular, is among OpenAI’s earliest and biggest supporters and has an agreement with the startup to use its artificial intelligence models.
That same year, OpenAI employees saw Pentagon officials walking through the company’s San Francisco offices, the report said, citing sources.
OpenAI then went on to update its blanket ban on military use cases in January 2024. In December of that year, the company also announced a partnership with Anduril to develop and deploy AI models for “national security missions.”
“Microsoft has a product called Azure OpenAI Service that will be available to the US Government in 2023 and is subject to the Microsoft terms of service,” Microsoft spokesman Frank Shaw told WIRED.
The company did not clarify whether Azure OpenAI would be available to the Pentagon, but noted that the service is not certified for “top secret” government workloads until 2025.
Meanwhile, OpenAI spokesperson Liz Bourgeois told the publication: “Artificial intelligence already plays an important role in national security, and we believe it is important to have a seat at the table to ensure it is deployed safely and responsibly.”
“We have been transparent with our employees in approaching this work, providing regular updates and dedicated channels where teams can ask questions and communicate directly with our national security team.” added
Since OpenAI’s deal with the Pentagon was announced last week, the company has faced mounting criticism from both inside and outside the startup.
The report notes that the Pentagon deal divided OpenAI employees, with some employees publicly voicing their concerns.
“The biggest losers in this whole process have been ordinary people and civilians in conflict zones,” Sarah Shoker, former head of OpenAI’s geopolitics team, said in a Substack post last week. “Our ability to understand the effects of military AI in warfare is and will be severely hampered by layers of obscurity caused by technical design and policy. These are black boxes through and through.”
“The biggest losers in all of this are ordinary people and civilians in conflict zones… Our ability to understand the effects of military AI in warfare is and will be severely hampered by layers of obfuscation caused by technical design and policy. They are black boxes through and through,” Sarah Shoker, former head of OpenAI’s geopolitics team, wrote in a post on Sustack last week.
Meanwhile, in an internal meeting with employees, CEO Sam Altman said he could not call on how the company’s defense department would use artificial intelligence tools. Altman also stated that the company is interested in selling its artificial intelligence models to NATO.
