Pentagon Bets on ChatGPT and Grok to Speed Up the Force

Pentagon Bets on ChatGPT and Grok to Speed Up the Force
The Pentagon sees a broader menu of approved AI tools as a productivity and security advantage, while its own counterintelligence officials warn that unvetted “shadow AI” can open dangerous paths for data leaks.

Nine months after launching GenAI.mil, the Department of War said it had onboarded more than 1.7 million of its 3 million-plus personnel to the generative-AI platform. On Aug. 31, it expanded that portal with Starshield AI’s Grok for Government and OpenAI’s ChatGPT Mil, adding to the frontier-model options already available to military users.

Both services were accredited to handle controlled unclassified information at impact level five — material that is not classified but still requires security safeguards. The rollout follows earlier Pentagon partnerships with OpenAI for classified-network models and with xAI to support personnel handling controlled unclassified information.

The department’s case is operational: Grok for Government is intended for tasks ranging from logistics and supply-chain work to acquisition-market research. It said the tool would deliver “immediate productivity gains, stronger knowledge continuity, and more secure and efficient collaboration.” Its features include adaptive reasoning, persistent projects and reusable playbooks.

ChatGPT Mil, meanwhile, is aimed at document-heavy unclassified work — planning, logistics, administration and policy. The Pentagon described it as bringing “a familiar commercial experience into the Department’s secure environment, tailored to warfighter needs.” Officials expect it to speed routine work so personnel can focus on more consequential Joint Force projects.

The strategy is not simply to deploy more AI, but to avoid reliance on one vendor. Yet the backdrop is a June warning from the Defense Counterintelligence and Security Agency that unauthorized commercial models and unreviewed embedded AI features create “a massive, unmonitored attack surface” where risks can outrun safeguards.

Continue reading https://foxvector.com/stories/01a05cf9-a846-0b7d-72ce-02ac08e5a377