More than an IT system in the government: The work divide challenges in human-AI coworking context

The study examines the collaboration between AI and humans in Taiwanese government agencies and the reasons for different modes of collaboration. Through interviews, four distinct patterns were identified based on the division of labor and the level of influence on citizens' rights. In cases where citizens' rights are minimally affected and the division of labor is independent, AI serves as a low-risk tool for routine tasks. For applications significantly affecting citizens' rights, especially those with mature AI technologies, government agencies tend to adopt a conservative approach, using AI as a reference or support rather than a primary decision-maker. The analysis also reveals instances where AI assists in tasks beyond its current capabilities, emphasizing the importance of Human-AI collaboration in resolving complex issues. Overall, the role of AI in Human-AI collaboration transcends traditional categorizations, showcasing nuanced relationships and varying degrees of collaboration across different contexts. Additionally, the study draws insights from past literature on ICTs' impact on bureaucratic systems, highlighting the evolving roles of public servants and potential challenges arising from technological advancements.