August 28, 2025
Subscribe and listen to the Fed Gov Today Podcast anytime on Apple Podcasts, Spotify, or at FedGovToday.com.
The federal government is experiencing a dramatic surge in the use of generative artificial intelligence (GenAI). According to new findings from the Government Accountability Office (GAO), agencies reported 282 GenAI use cases in 2024—up from just 32 the year before. “Overall, the use of generative AI, or GenAI, has escalated rapidly since 2023,” says Candice Wright, Director of Science, Technology Assessment, and Analytics at GAO. “It grew from 32 to now 282 use cases. So basically, a ninefold increase.”
Wright and her team examined how 11 agencies are experimenting with the technology by reviewing their use case inventories. They also found that AI adoption more broadly is booming. “The AI use cases had nearly doubled from 2023 to 2024,” she explains, “going from 571 to well over 1,000, or almost 1,100 use cases. So quite a lot of activity, lots of interest in exploring this technology in government.”
Some of the most dramatic growth is happening at the Department of Health and Human Services (HHS). “HHS went from seven use cases during 2023 to 116 in 2024,” Wright says. Much of that activity is in health and medical applications. She notes the agency is using AI to “extract information from lots of publications” and to “identify various disease outbreaks, like polio, for
But alongside the enthusiasm, agencies are confronting real challenges. Wright says one of the top concerns is “complying with existing federal policies and guidance.” Because GenAI evolves so quickly, “policies can [struggle to] keep pace with the technology itself.” She explains that agencies are particularly focused on privacy risks: “There’s certainly a unique risk when we talk about generative AI. We’re talking about creating new content…and so there’s just a concern, especially around privacy, whether or not the data that are being used to train these models and then the outcomes…are keeping pace with the way in which the technology operates.”
Resources are another sticking point. “One of the things that I would say comes at a really great concern for the agency is making sure that they have the advanced computing infrastructure that’s needed to support generative AI,” Wright says. That infrastructure is costly, and many agencies must weigh GenAI investments against other modernization priorities.
Finally, Wright points to the ongoing challenge of updating policies. “Officials at 10 of 12 selected agencies said existing federal policy…could present obstacles to adoption,” she explains. At the same time, the rapid evolution of GenAI “can complicate establishment of generative AI policies and practices.”
For Wright, the solution is collaboration. “When you think about AI more broadly, and certainly GenAI, it’s really a team sport. It’s an all hands on deck, lots of different perspectives that need to weigh in.” That means bringing together developers, legal experts, privacy officers, procurement specialists, and human capital leaders to shape governance.
Wright also emphasizes the importance of frameworks. “Back in 2020 we actually issued an AI accountability framework, really to guide responsible development and use of AI,” she says, noting that agencies continue to use GAO’s work alongside new executive orders, OMB guidance, and the NIST risk management framework.
Looking ahead, Wright encourages agencies to remain deliberate. “It’s really important…to take a look and say, okay, what problem is it that we’re trying to solve for, and is the technology the best or only means to do that?” she says. The federal government’s rapid expansion of GenAI shows both the promise and the responsibility of adopting powerful new tools in public service.
You can read GAO’s full report here