Who Is AI Replacing? The Impact of Generative AI on Federal Government Employment

• In Event: Insights on Artificial Intelligence in Government

Abstract

The rapid advancement of generative artificial intelligence (AI) has the potential to either complement knowledge workers by enhancing productivity or to replace them entirely. Recent empirical studies have documented that the adverse employment effects of AI adoption are disproportionately borne by workers in manufacturing, low- and middle-skill roles, and non-STEM occupations across U.S. commuting zones (Huang 2024). At the individual level, freelancers in highly AI-exposed professions—such as writing, proofreading, and copy-editing—are experiencing notable declines in both employment opportunities and earnings (Hui, Reshef, and Zhou 2024).

While AI’s labor market implications have been extensively studied in the private sector, its impact on public sector employment remains relatively underexplored. A recent survey of U.S. government employees highlights this uncertainty: while a majority of respondents believe AI will compete with human labor, opinions on job displacement remain divided (Ahn and Chen 2022). Specifically, 40% of respondents expressed concerns or uncertainty about AI’s potential to displace jobs, while 52% believed that AI could lead to the creation of new roles within government agencies. These mixed perceptions point to a fundamental and pressing question: Is generative AI more likely to replace or create jobs in the public sector?

This study employs a continuous treatment Difference-in-Differences design to estimate the impact of generative AI—such as ChatGPT—on federal government employment. I construct a treatment measure by linking federal job titles to an occupational AI exposure index and treating the emergence of generative AI tools as a policy shock. This approach enables a causal assessment of how varying levels of AI exposure across occupations influence federal employment outcomes over time.

This study offers two primary contributions. First, it fills a critical empirical gap by examining the effects of generative AI within the federal government—an area largely overlooked in existing research focused on the private sector. Second, it situates these effects within a broader historical context: technological change has long reshaped public employment, often disproportionately impacting less-skilled and older workers, who face greater barriers to reemployment after displacement (Aaronson and Housinger 1999). These findings highlight the importance of proactive policy interventions—particularly up-skilling and re-skilling programs—to mitigate AI’s adverse effects support a more inclusive transition to an AI-augmented public workforce (Soueidan and Shoghari 2024).

References

Aaronson, Daniel, and Kenneth Housinger. 1999. “The Impact of Technology on Displacement and Reemployment.” Economic Perspectives-Federal Reserve Bank of Chicago 23:14–30. 

Ahn, Michael J., and Yu-Che Chen. 2022. “Digital Transformation toward AI-Augmented Public Administration: The Perception of Government Employees and the Willingness to Use AI in Government.” Government Information Quarterly 39 (2): 101664. 

Huang, Yueling. 2024. “The Labor Market Impact of Artificial Intelligence: Evidence from US Regions.” IMF Working Papers 2024 (199): 1. 

Hui, Xiang, Oren Reshef, and Luofeng Zhou. 2024. “The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market.” Organization Science 35 (6): 1977–89. 

Soueidan, Mohamad Hasan, and Rodwan Shoghari. 2024. “The Impact of Artificial Intelligence on Job Loss: Risks for Governments.” Technium Soc. Sci. J. 57:206. 

Author

• Jie Tao, Tarleton State University
  Presenting Author