KISDI Finds Slower Low-Wage Employment Growth and Widening Wage Inequality in Industries with High AI Exposure (detail)

※ URL(Korean): https://www.kisdi.re.kr/bbs/view.do?bbsSn=115120&key=m2101113055776&pageIndex=1&sc=&sw=

KISDI Premium Report (26-7): AI Exposure and Wage Inequality: An Industry-Level Empirical Analysis

KISDI Finds Slower Low-Wage Employment Growth and Widening Wage Inequality in Industries with High AI Exposure

▲ Within-industry wage inequality tends to widen at the extremes of the wage distribution as AI exposure increases
▲ Wage gaps between industries widen as wage premiums increase in industries with high AI exposure
▲ Report recommends industry-specific policy responses and measures to strengthen labor market adaptability among low-wage workers

The Korea Information Society Development Institute (KISDI, President Sangkyu Rhee) recently published *KISDI Premium Report (26-7)*, titled "AI Exposure and Wage Inequality: An Industry-Level Empirical Analysis." The report empirically examines the relationship between AI exposure and wage inequality, focusing on changes within industries and wage gaps between industries. It also discusses the need for policy responses.

As AI becomes capable of performing not only routine and repetitive tasks but also non-routine and cognitive tasks, concerns about labor displacement have extended to high-skilled occupations traditionally considered areas of comparative advantage for human labor. However, the net effect of AI on employment remains uncertain. Even without substantial changes in total employment, AI may affect wage levels and distribution. Empirical evidence on the relationship between AI and wage inequality remains limited, and existing findings are inconsistent.

Against this background, the report analyzes changes between 2013, before the widespread adoption of AI, and 2023. Using the Regional Employment Survey conducted by Statistics Korea, the analysis covers 217 industries at the detailed classification level for which wage distributions can be calculated.

The results show that industries with higher AI exposure tend to experience greater within-industry wage inequality, particularly at the extremes of the wage distribution. In the model including all control variables, a 0.1-point higher AI exposure score is associated with an 8.13 percentage point higher growth rate in the P90/P10 wage ratio, which measures the wage gap between workers at the 90th and 10th percentiles.

By contrast, no statistically significant effects are found for the P80/P20 ratio or ratios based on median wages. This suggests that the effects of AI exposure may be concentrated at the extremes of the wage distribution rather than spread across the distribution as a whole.

In terms of employment composition, changes in industries with higher AI exposure are concentrated in low-wage jobs rather than overall employment. A 0.1-point higher AI exposure score is associated with a 16.9 percentage point lower employment growth rate among low-wage workers. No statistically significant changes are found in the growth rates of total employment or employment among high-wage workers. The share of low-wage workers within industries is also 2.616 percentage points lower.

Regarding between-industry wage gaps, industries with higher AI exposure show significantly higher average wage growth. Given that these industries already had higher average wage levels in 2013, the findings suggest that wage premiums in high-AI-exposure industries may be increasing, potentially widening wage gaps between industries according to their levels of AI exposure. The analysis therefore indicates that assessments of AI-related labor market changes should consider both within-industry wage distributions and between-industry wage disparities.

The report recommends differentiated policy approaches that reflect industry-specific characteristics, as AI exposure levels, employment structures, and wage distributions vary across industries. In particular, given the relatively slower growth of low-wage employment in industries with high AI exposure, the report emphasizes the importance of proactive support for job transitions and improvements in AI-related skills and literacy among low-wage workers.

It also recommends establishing an analytical framework to continuously monitor risks to low-wage jobs and changes in wage distributions across industries.

Jieun Choi, Research Fellow at KISDI, noted that AI's effects on the labor market may emerge in qualitative aspects of employment, including wage levels and distributions, even without substantial changes in employment levels. She explained that analyses focusing primarily on employment indicators may therefore fail to fully capture these distributional effects.

She also emphasized the need for a monitoring framework to track industry-level AI exposure and changes in wage distributions across percentiles, as the effects of AI on wage distribution may evolve over time.

KISDI Premium Report is available for download from the KISDI website (www.kisdi.re.kr).

KISDI News

KISDI Finds Slower Low-Wage Employment Growth and Widening Wage Inequality in Industries with High AI Exposure