The Department of War launched Grok for Government, an artificial intelligence tool developed by Elon Musk’s defense-focused subsidiary Starshield AI, on the newly created GenAI.mil platform for military personnel, on Sept. 1.
The deployment makes a commercial generative AI product available to service members for both classified and unclassified workflows, marking a significant expansion of AI tooling inside the Pentagon. The platform’s launch under a brand-new government domain establishes dedicated military infrastructure for generative AI operations.
Starshield AI is the defense-oriented arm of Musk’s technology empire, which also includes SpaceX, a major Pentagon launch and satellite contractor. The decision to deploy Starshield’s product on a military platform raises questions about procurement processes and potential conflicts of interest given Musk’s existing defense contracts and his relationship with the current administration.
The Department of War did not detail the competitive bidding process, if any, used to select Grok for the platform. Data security protocols governing how the AI system handles classified military information were also not described in the initial announcement.
Generative AI tools like Grok can draft documents, analyze data, summarize intelligence reports and assist with planning — functions that could accelerate military decision-making but also introduce risks if the systems produce inaccurate or biased outputs.
The deployment comes as the Pentagon has been rapidly accelerating its adoption of commercial AI products, a shift from previous administrations’ more cautious approach to integrating private-sector technology into defense operations.
Military AI deployments have drawn scrutiny from lawmakers on the Senate Armed Services Committee, who have pressed for transparency on how the Pentagon evaluates and secures commercial AI tools before giving them access to sensitive defense information.
The Government Accountability Office has previously flagged concerns about the speed at which federal agencies adopt AI systems without completing required security assessments and testing protocols.