From routine jobs to algorithmic labor: a digital labor history of Generative AI and shifts in employer hiring demand in China, 2016–2024

Abstract

This article treats online job advertisements as a digital archive of how firms classify skills, credentials, and tasks, thereby reconstructing the early evolution of employer demand in the context of generative artificial intelligence (GenAI) in China. Drawing on more than 13 million job advertisements posted on major Chinese recruitment platforms between 2016 and 2024 and matching them with records of listed firms, the study examines how firms that had participated more extensively in AI-related recruitment prior to the public emergence of GenAI adjusted their advertised wage, educational, and task requirements after ChatGPT brought GenAI into public awareness in November 2022. The results show that firms with greater prior involvement in AI-related recruitment exhibit more pronounced polarization in their recruitment demand. In terms of educational and wage requirements, the human capital structure of recruitment demand displays a K-shaped pattern: the share of middle-level positions declines, while the shares of high- and low-level positions expand relatively. The empirical results from the mechanism analysis indicate that this pattern arises because GenAI increases the cognitive complexity of work while reducing its operational complexity. Consequently, firms increasingly concentrate their recruitment demand at the two ends of the labor market.