The US Department of Defense has launched customized versions of OpenAI's ChatGPT and xAI's Grok on its secure GenAI.mil portal, expanding access to 3 million military and civilian personnel. The tools, ChatGPT Mil and Grok for Government, are designed for routine unclassified tasks and operate under strict security protocols to protect sensitive data. The rollout highlights the Pentagon's growing AI adoption and the ongoing exclusion of Anthropic's Claude due to a safety dispute.
GenAI.mil platform expansion
- ▪More than 1.7 million unique users out of the Department of Defense's 3 million personnel have already used the GenAI.mil platform.
- ▪The GenAI.mil platform initially launched offering Google's Gemini model before adding ChatGPT Mil and Grok for Government.
- ▪The US Department of Defense launched customized versions of OpenAI's ChatGPT, named ChatGPT Mil, and xAI's Grok, named Grok for Government, on its secure GenAI.mil portal on August 31, 2026.
- ▪The Department of Defense has secured artificial intelligence agreements with other technology companies, including Amazon Web Services, Microsoft, Nvidia, and Reflection AI.
- ▪The GenAI.mil portal, which launched on December 9, 2025, provides 3 million military and civilian personnel with access to commercial frontier artificial intelligence models.
ChatGPT Mil capabilities
- ▪Grok for Government, developed by SpaceX's Starshield AI using xAI's Grok, is designed to assist military personnel with productivity, knowledge continuity, and secure collaboration.
- ▪The capabilities of ChatGPT Mil will initially focus on chat, files, projects, and custom GPTs, offering an experience familiar to commercial ChatGPT users.
- ▪Work to create a version of ChatGPT for controlled unclassified information began in February 2026, with deployment initially anticipated for early July 2026.
Grok for Government deployment
- ▪The military versions of ChatGPT and Grok on GenAI.mil are engineered to operate in a separate environment from commercial models to protect controlled unclassified information and prevent data collection.
- ▪The Department of Defense stated that Grok for Government will support operational contexts ranging from market research analysis for acquisitions to supply-chain management for logisticians.
Controlled unclassified information security
- ▪Anthropic's Claude model is absent from the GenAI.mil platform following a dispute where Anthropic refused to grant the Pentagon unrestricted use of its artificial intelligence tools without safety guardrails.
- ▪OpenAI and Department of Defense officials stated that the artificial intelligence models on the GenAI.mil platform are not intended to be used for combat operations.
- ▪Preparing a frontier artificial intelligence model for GenAI.mil requires passing security guardrails from both the Defense Information Systems Agency and the National Security Agency.
Story comments
Loading comments…