Generative AI and Uneven Productivity Gains: The Political Economy of Sectoral Asymmetry in South Korea

ABSTRACT

  The diffusion of generative artificial intelligence (GenAI) raises questions not only about productivity but also about inequality, labor politics, and the future of development models. This paper examines South Korea – an economy long reliant on export-oriented manufacturing – to assess whether early GenAI adoption is associated with emerging differences in productivity across industries. Using quarterly data for 18 industries from 2013 to 2025, the analysis shows that industries with greater exposure to AI-augmentable tasks, many of which are in services, experienced short-run productivity increases after 2023, while routine-intensive manufacturing displayed limited change. These uneven outcomes extend task-based accounts of technological change to the industry level and suggest emerging sectoral differences in GenAI adoption. The paper contributes by providing early national evidence on sector-level patterns of GenAI exposure and by situating these patterns within debates about South Korea’s manufacturing-centered growth model. The findings point to the policy challenges associated with uneven technological adoption and to their implications for industrial strategy and labor relations.

2026년 2월
저널: Journal of Information Technology & Politics (SSCI, JCR 랭킹 Q1)
논문 타이틀: Generative AI and Uneven Productivity Gains: The Political Economy of Sectoral Asymmetry in South Korea