From Understanding to Collaboration: Unpacking the Black Box of Collaboration Between Street-Level Bureaucrats and Artificial Intelligence

Street-level bureaucrats (SLBs) increasingly collaborate with artificial intelligence (AI), yet research on human-AI collaboration (HAIC) lacks evidence from public organizations. This study examines the association between SLBs’ understanding of AI and their AI collaboration, based on a department-level census of 175 SLBs in Hangzhou, where SLB-AI collaboration occupies a leading position in China. The key findings are as follows. (1) Based on respondents’ evaluations, SLB-AI collaboration is perceived to outperform both SLB-only and AI-only modes in work efficiency, but it does not significantly surpass SLB-only mode in emotional competence. (2) SLBs’ understanding of AI is positively associated with collaboration and serves as an enabling condition for it. (3) The relationship between understanding and collaboration is significant under low-to-moderate work stress, but not under high work stress. These findings open the black box of SLB-AI collaboration in public organizations, highlighting the need for human-centered SLB-AI discretion and bureaucratic systems.