The Pentagon Wants to Raw Dog the Latest AI Models on Classified Systems
The Pentagon is looking to expand its use of artificial intelligence across both unclassified and classified networks, but negotiations with major AI companies have hit a sticking point.
Defense officials want access to the most advanced models without any usage restrictions or heavy guardrails. According to Reuters, military officials argue they should be allowed to deploy AI however they see fit, as long as it complies with U.S. law.
The push comes as OpenAI announced Monday that it has made a customized version of ChatGPT available through the War Department’s AI platform, GenAI.mil. The platform, which launched in December, is used by roughly 3 million civilian and military personnel and already includes tailored versions of tools from xAI and Google’s Gemini.
“We are pushing all of our chips in on artificial intelligence as a fighting force. The Department is tapping into America’s commercial genius, and we’re embedding generative AI into our daily battle rhythm,” Secretary of War Pete Hegseth said in a press release about the platform. “AI tools present boundless opportunities to increase efficiency, and we are thrilled to witness AI’s future positive impact across the War Department.”
OpenAI’s version of ChatGPT on the platform is designed to help with day-to-day tasks like summarizing policy documents, drafting reports, and assisting with research. But Reuters reports that Pentagon officials are pushing to roll out AI systems across all classification levels, potentially opening the door to more sensitive applications like mission planning or weapons targeting.
Currently, Anthropic’s models are available in select classified settings through third-party providers, but with significant usage restrictions. Reuters reports that Anthropic executives have told military officials they do not want their systems used for autonomous weapons targeting or domestic surveillance.
Meanwhile, Semafor reports that Anthropic has not agreed to allow its models to be used for “all lawful uses.” As of now, its tools are not currently available on GenAI.mil.
The negotiations leave AI companies walking a delicate tightrope. On one side, there are employees who oppose military use of their systems and fear it will make it hard to recruit future employees. On the other side is the Pentagon, which represents a massive customer and a powerful political force. Semafor reported that Anthropic’s stance has “drawn ire from the Pentagon and the White House.”
At the same time, some OpenAI employees have expressed concerns about giving competitors an advantage by stepping back from defense work, according to Semafor.
