Despite the spread of generative artificial intelligence, the number of workers in occupations with high AI exposure has grown by nearly 250,000 over the past three years. However, more than 80% of that increase was concentrated in the Seoul metropolitan area, and employment among workers in their 20s declined across all regions of South Korea — raising concerns that AI could widen labor market disparities between regions and generations.
According to the “BOK Issue Note: AI and Regional Labor Markets — Risks of Widening Regional Gaps and New Opportunities” released on the 15th by the Bank of Korea’s regional economics team, jobs with high generative AI exposure increased by 247,000 in the three years since generative AI began spreading in earnest in 2023. Of that total, the Seoul metropolitan area accounted for 199,000, or 80.8%.
Occupations with high generative AI exposure are office, sales, and professional roles that involve substantial document preparation and information processing. Representative examples include accounting and bookkeeping, planning and marketing office jobs, telesales positions, and data and software specialists. In this study, the Bank of Korea categorized AI into three types — generative, agentic, and physical — and calculated AI exposure by occupation and region. AI exposure refers to the potential for AI utilization and automation across the various tasks that make up a job.
In occupations with high agentic AI exposure, such as managerial roles, employment rose by 105,000 last year, with the Seoul metropolitan area accounting for 88.3% (93,000) of that increase. In the Seoul metro area, employment growth rose proportionally with higher AI exposure, but this pattern was not observed outside the capital region.
Employment Grows, but Mismatch Deepens
Contrary to concerns that AI could rapidly replace jobs, employment has actually increased in occupations with high AI exposure. The researchers explained: “Service-sector companies in particular may be using AI to enhance information-processing capabilities and expand markets rather than to replace workers. Because AI is still automating only part of job functions, workers appear to be maintaining employment by focusing on remaining tasks.”
Jeong Min-su, head of the Bank of Korea’s regional economic research team, said: “The fact that employment growth has been strongly concentrated in the Seoul metropolitan area is likely attributable in large part to regional disparities in AI utilization and IT human capital.”
Even as jobs increased, companies in the Seoul metro area still reported feeling greater labor shortages than those in other regions. Signs of wage increases also emerged primarily in the capital region. The analysis suggests that while demand has risen for workers with expertise, experience, seniority, and social skills following AI’s spread, the supply of such workers has not kept pace, creating a mismatch.
Workers in Their 20s See Employment Decline
Workers in their 20s bucked the trend seen across all other age groups. Employment in occupations with high generative and agentic AI exposure declined in both the Seoul metropolitan area and other regions. The researchers attributed this to the fact that younger workers, who have had less time to build expertise and experience, are relatively more likely to find themselves competing with tasks that AI can handle.
Jeong said: “For young workers, employment in high AI-exposure occupations declined uniformly across both the Seoul metro area and other regions. This is because university education still remains focused on traditional knowledge and information accumulation.”
Risks of Widening Regional Gaps and Opportunity Factors
The Bank of Korea warned that if these trends continue, AI could further widen regional labor market disparities over the long term. Knowledge-intensive services and semiconductor manufacturing — industries with high AI exposure — are already concentrated in the Seoul metropolitan area, and because AI’s productivity gains grow when combined with skilled labor, AI investment and related talent could become even more concentrated in the capital region. In fact, “AI readiness” scores evaluating regional AI utilization capabilities were higher in Seoul and Gyeonggi Province than in other regions.
The report also raised the possibility that employment outside the Seoul metro area could suffer a greater shock if physical AI spreads in earnest. Industries with high physical AI exposure — where machines and robots directly perform tasks, such as manufacturing, transportation, lodging, and food services — are relatively more concentrated outside the capital region.
However, AI also has the potential to narrow regional gaps. If an “equalizing effect” emerges — where workers with less experience and skill see greater productivity gains from AI assistance — productivity in regions with relatively weaker human capital could improve more rapidly.
The growing value of on-site knowledge in the development of physical AI is another opportunity for regional areas. If regional manufacturing plants take on research and development functions such as product development and process improvement, rather than remaining simple production bases, regional manufacturing could transition into knowledge-intensive industries.
To address employment polarization in the AI era, the Bank of Korea proposed strengthening customized AI training systems within regions, fostering the convergence of regional core manufacturing with knowledge services and startups, and expanding investment in “regional AI hubs” at anchor universities.
Jeong said: “To specialize regional service industries, we need to strengthen AI education systems while fostering knowledge services and startups linked to each region’s core manufacturing industries. There is also a need to expand investment so that regional anchor universities can serve as ‘AI hubs’ that concentrate AI talent and innovation capacity.”