Implications of artificial intelligence and robots for employment and labor productivity : firm-level evidence from the Republic of Korea

We examine the implications of robots and artificial intelligence (AI) for employment and productivity, using a rich firm-level database from the Survey of Business Activities provided by Statistics Korea. While previous studies have explored the effects of robots and AI separately, we investigate their effects jointly within a unified framework. We deploy propensity score matching to control for firm characteristics, enabling a potential causal interpretation of the differential impacts of robots and AI. We find that the patterns of adopting robots and AI differ significantly across industries. Additionally, although the overall share of firms adopting robots is larger, AI adoption is more concentrated among bigger firms. Our main finding is that, while adopting robots and adopting AI both increase employment, only adopting AI improves labor productivity. However, such productivity gains are accompanied by a decrease in the labor share of income, suggesting a potential shift in value distribution favoring capital income. Furthermore, we find that the immediate impact of adopting both robots and AI is an increase in temporary but not permanent employment. Finally, there is no evidence that firms adopting both robots and AI improve their labor productivity, potentially reflecting a lack of synergy.