The Pentagon has launched custom versions of OpenAI's ChatGPT and xAI's Grok on its GenAI.mil platform, giving about 3 million military and civilian personnel access to defense-tailored AI tools. The move bypasses Anthropic, which declined to provide its models without safety guardrails and was labeled a supply-chain risk by the administration.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
The U.S. Department of Defense has deployed customized versions of ChatGPT and Grok on the GenAI.mil platform, covering approximately 3 million personnel, while excluding Anthropic due to safety guardrail disputes, sparking discussions on the boundaries of military AI applications.
On September 1, 2026, the Pentagon announced the launch of customized versions of OpenAI's ChatGPT and xAI's Grok on the GenAI.mil platform, named ChatGPT Mil and Grok for Government, respectively, aiming to provide generative AI tools to approximately 3 million military and civilian personnel. GenAI.mil is a secure portal launched last year that allows Department of Defense employees to use commercial AI models without transmitting sensitive data through consumer-grade channels.
GenAI.mil initially offered Google's Gemini, which has reportedly been adopted by over 1.7 million users, representing the majority of the DoD's 3 million personnel. The newly added ChatGPT Mil, developed through OpenAI's government program, supports routine non-classified administrative, logistical, and planning tasks, with an experience similar to the commercial version of ChatGPT.
Grok for Government is built on SpaceX's Starshield AI satellite network, and the DoD's description of it is broader, aiming to support various mission executions from acquisition research to supply chain logistics. This suggests that xAI's model may take on more complex responsibilities in military scenarios, while ChatGPT Mil focuses on routine administrative work.
Notably, the Pentagon also reached AI-related agreements with Amazon Web Services (AWS), Microsoft, Nvidia, and Reflection AI, indicating a diversified approach in the military AI landscape.
In this update, Anthropic's Claude model was explicitly excluded. According to reports, Anthropic had refused to allow the Pentagon unrestricted use of its models without safety guardrails, leading the government to flag it as a supply chain risk. Anthropic is contesting this designation in court.
This incident highlights the tension between AI companies' safety stances and government needs in military applications. Anthropic insists on safety guardrails, while the Pentagon seeks more flexible AI tools, ultimately choosing other providers.
This article's information primarily comes from The AI Insider's report, which is a secondary source. Details about user numbers, model capabilities, and disputes are based on that report and have not been independently confirmed by official sources. Anthropic's stance and litigation status are as reported and should be verified with official statements.
The Pentagon has expanded the scope of military AI applications by introducing ChatGPT and Grok, while excluding Anthropic due to safety disputes, reflecting the trade-off between safety and efficiency in military AI deployment.
Primary report
Primary source