On My Desk: Evolving AI Policy & the Public Administrator

The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.

By Grant E. Rissler
August 22, 2025

Among the recent articles published in Administrative Theory & Praxis (ATP) is Evolving AI policy and the public administrator. The article reviews key AI policy trends across three recent presidential administrations, takes stock of early AI policy foci in the second Trump administration and summarizes several key implications for public administrators. These implications, in turn, provide food for thought for public administrators on the front lines of change.    

If you are among those who regularly use, or even once tried out ChatGPT, you likely received an e-mail last week announcing the arrival of ChatGPT5. The new version received mixed reviews, but the release and immediate comparisons to prior versions or competitors, as well as short lists of prompts to use with the revised tool highlight how widespread Artificial Intelligence (AI) use has gone from novel to being a part of normal life in just a few short years. (The image above was generated with ChatGPT5 using the following prompt—”Generate an image that evokes multiple stakeholders and the use of AI by public administrators”)        

We’re also beginning to see how AI, along with other elements of what has been called the 4^th Industrial Revolution, will change the world of public administration. In their recent article “Evolving AI policy and the public administrator” in Administrative Theory & Praxis (ATP), M. Blair Thomas and Daniel Baracskay provide some key insights of AI related policy trends across recent U.S. presidencies and summarize several key implications for public administrators. Their article is one of two recently published articles in ATP that deal with AI—the other being Artificial intelligence policy, the Trump Administration and federalism by Mallinson, Azevedo, Best and Robles. With limited space, this column focuses on the article by Thomas and Baracskay. 

Insights from the Article

While noting that AI, or “machines capable of sophisticated (intelligent) information processing” reach as far back as the 1950’s in their development, the authors also highlight that the first national (Canadian) strategy related to AI was only developed in 2017 but that 60 countries have followed suit. One significant contribution of the article is to summarize key trends across the second Obama administration, the first Trump administration and the recent Biden administration. The resulting table on page 7 of the online article traces an arc from early observations that no universal definition of AI existed through an emphasis on growing security risks (Trump I) to a focus under the Biden administration on how AI-impacted outcomes could avoid bias and yield equitable results. 

The article also names several key areas that are or should receive attention in the current Trump administration:

• Geopolitical goals of cementing U.S. leadership in AI and maximizing sovereignty over AI processes
• Attention to how human-machine coproduction enabled by AI may shift or require multistakeholder involvement and the preferences across different agencies and administrations for whether AI applications are primarily built and sourced in-house or purchased from commercial contractors (giving tech giants a larger influence over governance).
• Balancing efforts to maximize government efficiency with other values, such as engagement with the public and the possible cascade effects as human-machine coproduction gives way to automation with less human oversight/input.
• Relatedly, the needs for strategies “that align AI with public needs and priorities, ensuring outcomes that serve the public interest rather than relying primarily on efficiency and private sector models.” This insight may be particularly relevant given recent reporting on AI assisted DOGE efforts to eliminate large swaths of regulation.

Relevance to Public Administrators      

In addition to tracing several key trends, the authors also highlight a range of implications for public administrators, who, they note, will “shape the everyday use of AI practice” in governance. First, administrators will manage in a work culture impacted by significant worker reductions and will need to adapt quickly to meet outcome demands with less staffing resources. Second, administrators will need to rethink their positions relative to AI as AI becomes “more of an intermediary, or even active participant in…the decision-making process.”

They conclude by arguing that “public administrators at all levels should venture to involve multiple stakeholders in the [different] stages” of AI’s integration into government efforts.   

Conclusion

For society as a whole and those who administer public organizations, the infusion of AI into myriad processes and outcomes will require rethinking many elements of governance and citizenship. In the face of such rapid change, this article by Thomas and Baracskay helps not only by foregrounding the role of public administrator, but by emphasizing the role administrators will play in shaping stakeholder relationships even as automation of specific tasks increases.        

---

Author: Grant Rissler is Assistant Professor of Organizational Studies at the School of Professional and Continuing Studies, University of Richmond (VA).  He serves on the editorial board of Administrative Theory & Praxis (ATP) and focuses his research on social equity and peacebuilding with particular interest in local government responsiveness to immigrants.  The “On My Desk” series of columns, begun in July 2024, intentionally highlights the insights of one or more articles published in ATP in relation to a current debate or event.  Grant can be reached at [email protected].

(No Ratings Yet)
Loading...