Pentagon quietly began testing OpenAI models in 2023 despite military ban: Report

OpenAI signed a deal last week with the US government to allow its AI models to be used by the military for classified purposes. However, a new report by WIRED notes that the US military began experimenting with OpenAI models in 2023, despite the company having a blanket ban on the military's access to its AI models.

Reportedly, in 2023, OpenAI employees discovered that the Pentagon had begun experimenting with their models via Azure OpenAI, a version of the OpenAI models offered by Microsoft. The report notes that at the time, Microsoft had been contracting with the Department of Defence for decades.

Notably, Microsoft was among the earliest and biggest backers of OpenAI and has an agreement with the startup to utilise its AI models.

While citing sources, the report notes that in the same year, OpenAI employees saw Pentagon officials walking through the company's San Francisco offices.

OpenAI then updated its blanket ban on military use cases in January 2024. In December that year, the company also announced a partnership with Anduril to develop and deploy AI models for “national security missions.”

Microsoft spokesperson Frank Shaw, in a statement to WIRED, said, “Microsoft has a product called the Azure OpenAI Service that became available to the US Government in 2023 and is subject to Microsoft's terms of service.”

The company did not clarify if it had made Azure OpenAI available to the Pentagon, but noted that the service was not approved for “top secret” government workloads until 2025.

Meanwhile, OpenAI spokesperson Liz Bourgeois told the publication, “AI is already playing a significant role in national security, and we believe it’s important to have a seat at the table to help ensure it’s deployed safely and responsibly.”

“We've been transparent with our employees as we’ve approached this work, providing regular updates and dedicated channels where teams can ask questions and engage directly with our national security team,” she added.

OpenAI staff divided over Pentagon deal:

Ever since OpenAI's deal with the Pentagon was announced last week, the company has faced increasing criticism from both within and outside the startup.

The report notes that the Pentagon deal has divided OpenAI employees, with some employees even publicly raising their concerns.

“The biggest losers in all of this are everyday people and civilians in conflict zones,” said Sarah Shoker, the former head of OpenAI’s geopolitics team, in a Substack post last week. “Our ability to understand the effects of military AI in war is and will be severely hindered due to layers of opacity caused by technical design and policy. It’s black boxes all the way down.”

Meanwhile, at an internal meeting with employees, CEO Sam Altman reportedly said that the company does not get to make the calls on how the Defence Department uses its AI tools. Altman also noted that he is interested in selling the company's AI models to NATO.