AI Acceptance and Bureaucrats' Self Determination: A Mixed Method Study in South Carolina

• In Event: AI Adoption and Institutional Governance

Abstract

This study examines the motivational drivers of AI acceptance among local public employees, addressing a critical gap in behavioral public administration. Grounded in Self-Determination Theory (SDT), we argue that employee willingness to adopt AI depends on whether these systems fulfill or threaten fundamental psychological needs for competence, relatedness, and autonomy.

We employ a mixed-methods approach to test these dynamics. First, we conducted interviews with city and county leaders in South Carolina to inform a conjoint experiment. A key contribution of this work is the extension of the traditional SDT framework; based on our qualitative findings, we incorporate transparency as a critical fourth dimension essential for public sector legitimacy. We then surveyed a diverse sample of 1,100 local government employees, presenting them with randomized scenarios that varied how an AI system supported these four psychological needs.

Our findings suggest that public employees are not inherently technophobic. Rather, resistance is highly contingent on the autonomy configuration of the tool. Respondents expressed a strong, significant preference for augmented intelligence, where AI flags issues but human staff retain final decision-making authority, over systems that independently approve or deny permits. Furthermore, transparency is a non-negotiable prerequisite. Systems that obscure decision criteria or "hide" the AI’s role significantly reduce employee acceptance.

Qualitative data complement these findings, revealing that "fear of the unknown" and concerns over algorithmic accuracy drive resistance more than a simple aversion to change. Ultimately, this research suggests that successful AI integration in the public sector must prioritize "human-in-the-loop" designs that empower practitioners, ensuring technological efficiency does not come at the cost of administrative discretion or accountability. Together, our findings suggest that the successful integration of AI in local government depends less on the sophistication of the technology and more on a design philosophy that empowers public employees and ensures the continued legitimacy of democratic decision-making.

Authors

• Dongfang Gaozhao, University of Dayton
  Presenting Author

• Jun Li, University of Illinois, Chicago
  Non-Presenting Co-Author