The Pentagon is using generative AI to draft the congressionally mandated reports it owes lawmakers each year, with Chief Technology Officer Emil Michael telling the Hudson Institute on June 12 that a single report that once consumed 200 staffing hours can now be produced in five. The tooling runs on GenAI.mil, the Department of Defense's bespoke platform built around Google Cloud's Gemini for Government, which has been available across all six military branches since December 2025.
Adoption has scaled fast. Michael said personnel using GenAI.mil rose from 80,000 in December 2025 to 1.5 million in June 2026 — roughly 43% of the department's 3.5 million-person workforce in six months. The Pentagon is making that case as it asks Congress to approve a $1.5 trillion budget for the 2027 fiscal year.
The reporting burden is real. The number of reports Congress requires from the Department of Defense climbed from just over 500 in 2000 to more than 1,400 by 2020, according to the US Government Accountability Office. Elizabeth Field, a former GAO senior executive director, told Federal News Network in 2023 that the Office of the Assistant Secretary of Defense for Legislative Affairs has to comb defense authorization statutes "almost line by line" to find new reporting requirements, a process that can take three to six months before drafting even begins.
Key facts
- 01Pentagon CTO Emil Michael says generative AI cut a congressional report from 200 staffing hours to 5.
- 02GenAI.mil users jumped from 80,000 in December 2025 to 1.5 million in June 2026, out of a 3.5M DoD workforce.
- 03Congressionally mandated DoD reports rose from 500 in 2000 to more than 1,400 by 2020.
- 04The Pentagon signed May 1 agreements with eight frontier AI companies, including OpenAI, Google, Microsoft, and SpaceX — but not Anthropic.
- 05The Department of Defense has requested a $1.5 trillion budget for fiscal 2027.
Michael's five-hour anecdote is not the only example officials have offered. At the Box Federal Summit on April 23, Deputy Assistant Secretary of Defense Jacob Glassman described telling a short-staffed team responsible for a congressionally mandated report to "use GenAI.mil, do the best you can." Glassman said the team came back a week later claiming the output was the best report the office had produced in five years. He did not identify the report.
Beyond formal reports to Congress, service members are using generative AI to draft personnel evaluations for non-commissioned and commissioned officers, generate commendation medal citations, and write counseling statements, according to a Small Wars Journal article cited in the original reporting.
“the best report we've written in the past five years.”— Jacob Glassman, Deputy Assistant Secretary of Defense for Science and Technology Foundations
The accuracy question is unresolved. The Pentagon has not described what review process it applies to AI-generated reports before they reach Congress. Other organizations have already hit the obvious wall: consulting firm KPMG pulled a report titled "Redefining excellence in the age of agentic AI" after the research group GPTZero identified case studies containing AI-generated errors and false claims. Congressional reports are a primary mechanism through which lawmakers track how the military spends taxpayer money, and undetected fabrications would degrade that oversight at the same moment the department is asking for its largest-ever budget.
The vendor stack reflects a broader federal procurement push. Multiple US tech companies signed agreements with the General Services Administration in 2025 to provide AI tools across federal agencies at discounted rates. On May 1, the Department of Defense announced contracts with eight frontier AI companies to deploy tools on classified networks for "lawful operational use": SpaceX, OpenAI, Google, Nvidia, Reflection AI, Microsoft, Amazon Web Services, and Oracle. Contract values have not been disclosed.
Anthropic is not on that list. The Trump administration excluded the company after it declined to allow Claude to be used without restriction for autonomous warfare and mass surveillance applications, per the source reporting. That exclusion is a notable competitive data point for a lab whose federal revenue ambitions have been a central part of its growth story, including the recent DOJ filings backing rival xAI's Grok on military grounds that AI Chat Daily covered last week.
The productivity math is straightforward and hard to argue with: a 40x reduction on a single high-friction task, multiplied across more than 1,400 annual reports, frees thousands of staffing hours for substantive work. If the Pentagon can build a verification layer that catches hallucinated citations and fabricated figures before reports reach the Hill, GenAI.mil becomes one of the clearest examples of large-organization AI deployment delivering measurable returns. If it cannot, the same tooling will quietly inject errors into the document trail Congress uses to authorize a $1.5 trillion budget — and the first time a fabricated statistic surfaces in a hearing, the political cost will land on Google, OpenAI, and every other vendor whose logo is on the platform.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
