The U.S. Department of Defense has expanded its arsenal of standard artificial intelligence tools: on the internal secure portal GenAI.mil, military and civilian employees of the department have gained access to ChatGPT from OpenAI and Grok from xAI. According to the department's press release, cited by Ukrainian and Russian media, the integration is being carried out in a specially isolated environment designed to automate routine processes without the risk of leaking classified data. This move comes amid a public conflict between the Pentagon and the company Anthropic, and effectively cements the defense department's course toward a multi-provider model of AI partnership.

The Secure GenAI.mil Environment and Its Audience

According to reports, the GenAI.mil portal has already established itself as an in-demand tool within the department: of approximately 3 million Department of Defense employees, more than 1.7 million unique users have used the platform. Prior to this expansion, only one AI tool was available on the portal — Gemini from Google. The addition of ChatGPT and Grok turns GenAI.mil into a multi-platform environment where the choice of model is determined by the nature of the task, rather than by a tie to a single vendor.

The Conflict with Anthropic as a Trigger for Change

The expansion of the partner list comes against the backdrop of a public dispute between the Pentagon and the developer of the Claude model, the company Anthropic. According to available information, the conflict arose because Anthropic refused to provide unlimited access to its model, insisting on retaining built-in safety guardrails. In response, the U.S. administration placed the company on the supply chain risk list, while Anthropic, in turn, is challenging this decision in court. It is this dispute, according to reports, that prompted the defense department to accelerate the integration of alternative AI systems.

ChatGPT Mil: Functionality and Scope of Application

The version of ChatGPT for the department was created under the OpenAI for Government program and replicates the familiar set of features for users: working with chats, files, projects, and configured tools. As stated in the materials, ChatGPT Mil will be used primarily for non-classified routine work — document analysis, administration, planning, and the development of internal rules. Thus, OpenAI gains access to a broad but administratively oriented segment of the Pentagon's tasks.

Grok from xAI in a Broader Military Context

The integration of Grok from Elon Musk's xAI, implemented through the Starshield AI satellite network, is positioned by the department in a broader and directly military context. The AI tool is intended to accelerate the execution of missions and improve the accuracy of actions in various operational conditions. In addition, Grok will be used for market analytics in procurement and supply chain management for military logistics, which extends the tool beyond office automation into the realm of operational and logistical support.

What This Means Overall

Taken together, the announcement shows that the Pentagon is building a risk-resilient generative AI ecosystem: instead of relying on a single provider, the department is distributing tasks among OpenAI, xAI, and other partners (the list also mentions Amazon Web Services, Microsoft, Nvidia, and Reflection AI). The key condition remains operation within the secure GenAI.mil environment, which separates non-classified automation from classified processes and minimizes the risk of data leaks.