AI street-level bureaucracy: deployment and preferences for public sector employment
通过离散选择实验和访谈，研究香港硕士生对AI街头官僚不同部署属性的偏好，发现他们更偏好AI决策自主性低、性能与人类相当、提供AI培训和技能补贴的公共部门工作。
Artificial intelligence (AI) is increasingly employed in street-level bureaucratic encounters, giving rise to AI street-level bureaucrats (AI SLBs). Yet how specific attributes of working alongside AI SLBs shape preferences for public sector employment remains poorly understood. We conducted a discrete choice experiment (n = 1,320 observations) and follow-up interviews (n = 40) with final-year master’s students in Hong Kong to examine tradeoffs among five AI SLB deployment attributes: AI SLBs’ deployment intensity, performance, decision-making autonomy, workplace AI training, and AI skills development subsidies. Results reveal that participants prefer public sector jobs featuring lower AI SLBs’ decision-making autonomy, AI performance equivalent to humans, provision of AI-related training, and higher subsidies for AI skills development. Deployment intensity was not a significant predictor. This study advances the literature on public sector employment preferences in the AI era by identifying which attributes of human-AI collaboration matter most to prospective workers. Given that the findings from this strategically selected group may not generalize to all public sector seekers, we advocate for further research examining a broader range of stakeholders involved in public sector employment.