The Effects of Artificial Intelligence and Robotics on Employment and Wages in Korean Manufacturing Firms

Abstract

This article analyzes the effects of two key automation technologies – artificial intelligence (AI) and robotics – on employment and wages in Korean manufacturing since the late 2010s. Drawing on firm-level data from the Survey of Business Activities and individual wage data from the Local Labor Force Survey, the analysis explores both firm- and worker-level impacts. Adoption of these technologies is concentrated in large firms within the electronics and automotive sectors. Robotics has been widely implemented, primarily for cost reduction, safety enhancement, and union avoidance, whereas AI adoption remains limited but is gradually expanding. The results reveal contrasting effects: AI adoption is associated with job creation and wage growth, while robotics tends to reduce both employment and wages – an outcome that diverges from findings in existing firm-level studies. These negative effects appear to stem from Korea’s institutional context, where automation – particularly robotics – is frequently employed to reduce labor costs rather than to enhance productivity, as well as from diminishing marginal returns in industries with long-standing automation. Importantly, these wage effects persist even when U.S.-based automation exposure measures are applied, suggesting broader applicability. However, the findings underscore that the economic impact of automation depends significantly on the motivations and strategies behind its adoption. In the case of Korean manufacturing, capital-biased automation driven by robotics has contributed mainly to labor displacement without generating substantial productivity gains, reflecting Acemoglu and Restrepo’s (2018) notion of ‘so-so automation.’