The Korea Employment Information Service published its quarterly Employment Issues on 31 August, themed on AI and the labour market. The PDF's own upload stamp is 09:55 KST — 00:55 UTC. It scores 415 occupations for AI exposure against the national occupational database and the employment-insurance records.
The numbers
19.1% of employment-insurance subscribers — 2,916,000 of 15,275,000 — sit in the high-exposure group. That group is younger than the workforce: mean age 40.2 against 45.7 overall. Medium-to-high exposure by age band runs 53.1% for workers in their thirties, 44.1% under thirty, 42.7% in their forties, 27.2% in their fifties and 19.2% at sixty and over. A separate establishment survey of 2,297 workplaces found 28.6% adoption, with 75.6% of adopters using generative AI.
Read the top of the ranking
The most exposed occupations are lawyers, then legal administrative staff, patent attorneys, life-science technicians and accountants. That ordering is the key to the whole report. If this were a displacement index, lawyers would not lead it. It is a task-overlap index — a measure of how much of an occupation's work a model can perform — and licensed professions score highest precisely because their work is documentary. Overlap is not substitution: the tasks a model can do and the tasks an employer is permitted to let it do are different sets, and the gap is widest exactly where licensing exists.
What the common framing gets wrong
Coverage ran "one in five Korean workers exposed to AI" and "half of thirty-somethings at risk," converting an overlap score into a risk forecast. The report's own analysis blocks that. It finds new employment-insurance entries fell across every exposure group — low as well as high — which is the pattern of a general hiring slowdown, not of AI-driven displacement. And it attributes the youth hiring decline mainly to the shrinking youth cohort, a demographic fact that would be operating with or without AI.
The one comparison that does point somewhere
The youth employment index from 2022 to 2025 sits at 86.6 for high-exposure occupations against 92.5 for low-exposure ones. That gap is the report's most suggestive finding, and it is a difference of about six points over three years — real, measurable, and considerably smaller than the headlines built on top of it.