The Defense Department is now using artificial intelligence platforms to draft reports to Congress, informing lawmakers on a variety of national security issues so they can set policy. (Department of Defense)
The Defense Department is increasingly using artificial intelligence to handle administrative work, including drafting congressionally mandated reports, as Pentagon officials push to expand the technology’s use across the military.
Emil Michael, the Pentagon’s chief technology officer, championed the use of AI for writing reports to lawmakers during a Hudson Institute forum in Washington, D.C., on June 12.
“I have to report to Congress every year on this thing,” Michael said at the forum. “Let me load all the papers on and have it draft me a congressional report that would otherwise take 200 hours of staffing time and do it in five hours. It’s just a matter of trying to catch up to, in this case, what’s basic in the commercial world.”
Michael brushed off questions about risks and the potential for AI hallucinations.
“I mean, there’s risk in everything, and we also offer training,” he said.
Pentagon officials draft hundreds of reports per year, and these documents inform lawmakers who generate policy on a variety of national security issues.
Defense Secretary Pete Hegseth, left, and Undersecretary of Defense for Research and Engineering Emil Michael tour autonomous systems at the Pentagon in Washington on July 16, 2025. The Defense Department is encouraging staff to use artificial intelligence to draft congressionally mandated reports. (Alexander Kubitza/U.S. Navy)
A Marine works at his computer at Marine Corps Base Camp Lejeune, N.C., Jan. 21, 2026. The Defense Department is encouraging staff to use artificial intelligence to draft congressionally mandated reports, as the number of required reports has tripled over the past two decades. (Brady V. Hathaway/U.S. Marine Corps)
The use of AI to draft congressional reports is part of a broader Pentagon effort to incorporate the technology into administrative, intelligence and warfighting functions.
In June, Michael said the number of Defense Department personnel using AI tools had grown from about 80,000 to more than 1.5 million over the past year, a more than 17-fold increase.
The Pentagon began using Google’s Gemini for government use on its custom GenAI.mil platform in December. The AI hub has since grown to include xAI’s Grok and OpenAI’s ChatGPT, according to Department of Defense statements.
Anthropic also was awarded a one-year, $200 million contract last year to develop AI infrastructure that would address national security challenges in “warfighting and enterprise domains,” the contract announcement from July 2025 states.
The growing use of AI for administrative tasks comes as the Pentagon’s congressional reporting requirements have expanded.
The number of reports the department submits to Congress increased threefold, from just over 500 in 2000 to nearly 1,500 in 2020, according to a Government Accountability Office report.
Michael is not the first Pentagon official to tout the use of AI in drafting congressional reports. Jacob Glassman, deputy assistant secretary of defense for science and technology foundations, told the story of a shorthanded team using AI to draft one such report during an April summit in Virginia, DefenseScoop reported.
“I said, ‘use GenAI.mil, do the best you can,’ because I’m not going to go back to Congress and say, ‘No, I’m not going to give you a report,’” he said, according to the outlet.
The team came back to him a week later and said it was the best report they had written in five years, DefenseScoop said. Glassman did not specify which report he was referring to.
Researchers say that while AI can improve efficiency, it can also provide inaccurate information. A BBC-led study published in October found that nearly half of the answers generated by major AI platforms about newsworthy events contained at least one significant error. Twenty percent of the responses included hallucinations or made-up information.
The study was coordinated by the European Broadcasting Union and included 22 public service media organizations in 18 countries working in 14 languages, the BBC reported.
The study examined answers provided by ChatGPT, Gemini, Microsoft Copilot and Perplexity.
The Pentagon declined an interview request this week from Stars and Stripes to discuss the use of AI in drafting reports and did not answer emailed questions about what policies exist to ensure accuracy.
