GenAI.mil, the Pentagon’s enterprise generative‑AI portal, logged 1.5 million daily users last week, according to Undersecretary of Defense for Research and Engineering Emil Michael. The figure represents almost half of the Department of Defense’s 3.5 million employees and marks a dramatic climb from the roughly 80,000 users who accessed the platform when it launched in December 2025.
Michael attributed the early lag to unclear guidance. "It wasn’t really clear where to go for it, what you could use it for, the rules were unclear," he said at a Hudson Institute event. The breakthrough arrived when the Pentagon integrated Google’s Gemini model onto its unclassified networks. Within weeks, daily log‑ins surged, and the platform soon added OpenAI’s ChatGPT and xAI’s Grok, giving every DoD worker a single portal to multiple commercial AI tools.
Most of the activity centers on back‑office work. Personnel draft job descriptions, summarize meeting notes, build budgets and, most notably, generate congressional reports. Michael cited a case where a task that once required 200 staff hours now finishes in five after loading source documents into the AI and letting it draft the narrative. "More and more people are like, ‘Oh my God, I could write a job description,’" he observed, noting the shift from simple queries to more sophisticated outputs.
The platform’s capabilities have moved beyond chat. In April, DoD staff used Gemini’s Agent Designer to spin up more than 100,000 semi‑autonomous AI agents in under five weeks. Operating at Impact Level 5—the highest classification for unclassified data—these agents draft after‑action reports, analyze operational datasets and review imagery, extending AI’s reach into mission‑critical analysis while staying within unclassified bounds.
Funding follows usage. The fiscal 2027 defense budget requests $54.6 billion for the Defence Autonomous Warfare Group, a steep jump from the previous year’s $13.4 billion allocation for AI and autonomy. The boost signals the Pentagon’s intent to embed AI across every function, from paperwork to battlefield decision‑making.
Not everyone is on board. More than 580 Google employees, including senior DeepMind researchers, signed a letter urging CEO Sundar Pichai to refuse classified military AI work, arguing that the company cannot monitor usage on air‑gapped networks. Google had already deployed Gemini to the Pentagon’s unclassified workforce and was negotiating classified access under “all lawful uses” terms. The Pentagon, meanwhile, signed classified AI agreements with seven firms—including Google, Microsoft, Nvidia and OpenAI—crafting terms that replace safety restrictions previously championed by Anthropic.
Unclassified adoption carries lower risk. Errors in a draft report, for instance, can be caught by human reviewers. The Pentagon has not released metrics on error rates or output accuracy, leaving analysts to wonder whether speed has outpaced quality assurance.
Michael framed the growth as organic. The department circulated case studies, and exposure to consumer AI tools outside of work helped staff navigate the new platform. Five of the six military branches now list GenAI.mil as their primary enterprise AI solution, making the rollout the fastest enterprise AI deployment in U.S. government history.
While the platform promises productivity gains, the Pentagon has yet to disclose how it will monitor hallucinations, factual errors or the broader implications of semi‑autonomous agents operating at scale. The story of GenAI.mil illustrates both the allure of rapid AI integration and the challenges of governing that power within a massive, bureaucratic organization.
This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.