Automation in South Korea: How AI and Robotics Are Reshaping Work and Wages

The study by ADB and Korea University examines the impact of AI and robotics on employment and productivity in South Korea, revealing that AI boosts productivity but reduces labor’s income share, while robots increase jobs without clear efficiency gains. Firms adopting both technologies see temporary employment growth but no productivity improvements, highlighting a lack of synergy.

The Asian Development Bank (ADB), in collaboration with Korea University, conducted an extensive firm-level study on how artificial intelligence (AI) and robotics are reshaping employment and labor productivity in South Korea. Using Statistics Korea’s Survey of Business Activities, the research employs propensity score matching (PSM) and two-way fixed-effects (TWFE) models to analyze the impact of automation at the firm level. South Korea, recognized as a global leader in automation, has over 1,000 robots per 10,000 employees, making it an ideal case study for understanding the labor market implications of AI and robotics.

The study finds that firms integrating AI and robots increase employment, but their effects on productivity and income distribution vary. AI adoption leads to higher labor productivity but also results in a decline in labor's share of income, meaning the economic gains from AI tend to benefit capital rather than workers. On the other hand, robots increase employment without necessarily improving productivity, raising concerns about their long-term efficiency. Interestingly, firms that adopt AI and robots experience only temporary employment growth, with no significant improvements in labor productivity, suggesting a lack of synergy between these technologies.

Industries That Benefit the Most from AI and Robotics

The adoption of AI and robotics differs significantly across industries. Robot adoption is more common in manufacturing, electricity, gas, and steam sectors, where automation can streamline production processes. In contrast, AI adoption is most prominent in information and communication, finance, and education, where decision-making and data processing play a critical role. Large firms are more inclined to invest in AI, while robot adoption is more evenly spread across firms of all sizes.

One of the key takeaways from the study is that robot adoption has little effect on permanent employment but temporarily increases temporary employment, indicating that firms are still experimenting with robotic technologies. AI adoption, on the other hand, leads to increases in both permanent and temporary employment, making it a more effective tool for workforce expansion.

Firms that adopt AI tend to have higher R&D intensity and lower capital intensity, suggesting that AI requires more knowledge and intellectual input rather than heavy machinery investment. In contrast, robot-adopting firms have lower R&D investment but a higher share of manufacturing workers, reflecting the traditional role of robotics in industrial automation. Another significant finding is that AI adoption appears to be driven by domestic factors, whereas robotics adoption is linked to global trade and supply chains.

Is There a Conflict Between AI and Robotics?

A surprising revelation from the research is that there is little correlation between AI and robot adoption, with a coefficient of just 0.2. This means that firms investing in AI are not necessarily the same ones investing in robotics, and vice versa. Even among firms that adopt both technologies, there is no clear evidence of productivity gains, challenging the assumption that AI and robotics naturally complement each other. This suggests that many businesses struggle to integrate both technologies effectively, leading to inefficiencies and underutilized automation potential.

This lack of synergy could be due to several factors, including the different skill sets required to manage AI and robots, as well as the difficulty in combining machine-learning-based decision-making with physical automation. If firms do not optimize these technologies together, they may fail to see significant improvements in productivity.

The Future of Employment in an Automated World

The policy implications of these findings are critical. The decline in the labor share of income among AI-adopting firms suggests that AI-driven productivity gains primarily benefit capital owners, potentially widening income inequality. Without proper interventions, AI could contribute to greater wealth concentration, leaving workers with fewer benefits. On the other hand, while robots create more jobs, their lack of productivity improvements raises concerns about whether firms are making the most of their automation investments.

To counter these challenges, policymakers must ensure that AI and robotic adoption leads to equitable economic growth. This includes restructuring tax policies, introducing AI-driven wage regulations, and supporting worker retraining programs. The research suggests that governments should focus on education and upskilling initiatives that align with technological advancements, preventing skill mismatches and ensuring that workers remain competitive in an AI-driven job market.

One of the most interesting findings is that firms adopting AI after 2019 experienced significant employment growth, while those that adopted AI earlier saw smaller or delayed effects. This could indicate that as AI technology evolves, its ability to generate jobs improves, making later adopters more successful in workforce expansion. Meanwhile, robot-adopting firms showed inconsistent effects on employment, with most only increasing temporary jobs in the short term.

Automation with a Human Touch: What Comes Next?

The findings make it clear that while AI and robotics are transforming the workplace, their effects are far from uniform. South Korea’s experience highlights the need for carefully designed automation strategies that balance employment growth with productivity improvements. Businesses need to approach automation strategically, ensuring that AI and robots do not simply replace workers but rather contribute to meaningful economic transformation.

At the same time, policymakers must address the inequalities created by AI-driven productivity gains by ensuring that technological progress benefits both capital and labor. This could involve creating AI taxation models that redistribute gains, offering incentives for firms that prioritize workforce training, and implementing regulations that encourage ethical AI use in employment.

Looking ahead, global studies should explore how different economic environments influence AI and robotic adoption outcomes. The South Korean case provides valuable insights, but expanding this research across multiple countries can help determine whether similar trends emerge in economies with different labor policies and technological infrastructures. Understanding these dynamics will be crucial for developing inclusive automation policies that drive productivity without compromising social equity.