In China’s rapidly developing AI industry, data annotation has traditionally been viewed as a low-cost, labor-intensive task (Wang et al., 2022; Yılmaz & Bostancı, 2025; Miceli et al., 2020). However, as China defines high-quality datasets as a decisive factor in its AI national competitiveness (National Development and Reform Commission, 2024), a growing trend has emerged where AI companies are increasingly hiring highly educated workers, specifically PhD students, to perform simple work, offering salaries nearly two hundred times higher than those paid to low-educated workers (Wang, 2023; Lin, 2023). This shift presents a paradox: Why are companies choosing to hire highly educated workers for tasks that are simple and could be performed by lower-wage labor?

This “irrational” puzzle is significant because it reveals how Chinese institutions are managing technological and economic uncertainties in the AI sector. As China continues to regulate its AI and digital economies, understanding labor practices in this industry provides insights into how the state adapts to broader industrial and technological challenges. Much of the existing literature on digital labor and AI training views these tasks as part of a rational, efficiency-driven system, this study offers a more powerful explanation from organizational lens.