According to an employment analysis published by the Korea Employment Information Service (KEIS), more than half of South Korean employment insurance subscribers in their 30s work in occupations heavily exposed to artificial intelligence disruption.
AI Exposure Concentrates in Prime Working-Age and Professional Roles
To determine occupational exposure, KEIS researchers combined the agency’s 2025 job incumbent survey data, employment insurance administrative databases, and international research findings. The resulting metrics show that out of 15,275,000 employment insurance subscribers registered last year, 19.1%—or approximately 2,916,000 workers—belong to high-AI-exposure groups, according to the KEIS report.
The data reveals a stark generational divide in technological risk. While the average age of all employment insurance subscribers stands at 45.7 years, workers in the high-AI-exposure category average 40.2 years old. Specifically, 53.1% of employment insurance subscribers in their 30s hold jobs classified within medium- to high-AI-exposure tracks, indicating that the workforce segments most vulnerable to automation are relatively young.
High-exposure groups encompass knowledge-intensive professions where text interpretation and quantitative analysis form core tasks. Lawyers, patent attorneys, accountants, judges, prosecutors, and industrial safety managers make up the bulk of this tier. Conversely, manual production, construction, and face-to-face service roles—such as electroplaters, plasterers, rice cake makers, and slaughterhouse workers—fall predominantly into low- or non-exposure brackets.
Shifts in Actual Employment Trends
The empirical data shows that theoretical exposure is already translating into measurable employment shifts. According to the KEIS analysis, job categories tied to accounting, tax assessment, and bookkeeping have experienced sustained declines in active employment insurance subscribers. Meanwhile, clerical positions such as customer service representatives, statisticians, and secretaries—classified under medium exposure—saw steady growth before reversing into a decline of a number of workers last year.
Yoon noted that while broader decreases in total employment insurance subscribers stem largely from supply-side factors like South Korea’s declining birth rate and aging population, policymakers must continuously monitor structural shifts in fields vulnerable to automation. Yoon emphasized the need to closely track employment dynamics among young adults, who face concentrated exposure, and within customer service sectors where generative AI and large language models (LLMs) are rapidly deploying.