Korea Employment Information Service publishes summer issue of 'Employment Issues'

19.1% of insured workers in high-exposure occupations, including journalists and legal clerks

Employment drop steeper among under-30s in high-exposure jobs

AI adoption rate at businesses stands at 28.6%; researchers call for job restructuring

A job placement center at a university in Seoul [Yonhap]
A job placement center at a university in Seoul [Yonhap]

About one in five employment insurance subscribers in South Korea works in an occupation with high exposure to AI technology — yet analysts say the recent hiring slowdown is driven more by a shrinking youth population than by AI displacing jobs.

The Korea Employment Information Service made the assessment Monday in the summer 2026 issue of its "Employment Issues" journal, examining how the spread of AI technology is affecting jobs and workplaces.

Linking data from the Korea Network for Occupations and Workers and the employment administration database, the agency calculated AI exposure levels across 415 occupations. Of the approximately 15.275 million employment insurance subscribers last year, 19.1 percent — about 2.916 million — fell into the high-exposure category. That share has held steady at 18 to 19 percent over the past five years.

High-exposure occupations include legal clerks, journalists and media professionals, government administration specialists, telemarketers, accounting and tax specialists, data professionals, finance and insurance professionals, business support clerks, and software developers. The number of insured workers among accounting, tax and appraisal specialists and bookkeeping clerks has continued to fall, raising concerns about a contraction in employment for roles centered on routine data processing.

Overall, however, the number of insured workers in high-exposure occupations continued to grow. While total employment insurance subscribers under 60 fell by 26,000 from the previous year, subscribers in high-exposure occupations increased. The agency noted that AI exposure scores reflect the degree of connection between an occupation and AI technology, and do not directly measure the likelihood of substitution or actual employment impact.

The employment slowdown appeared regardless of AI exposure level. Since 2023, when generative AI began spreading in earnest, the number of new employment insurance enrollees declined across high-, medium-, low- and no-exposure occupations alike.

A factor decomposition of employment insurance subscriber growth showed that insured workers aged 15 to 29 fell 4.6 percent last year. A decline in the youth population accounted for 2.4 percentage points of that drop, while a fall in the labor force participation rate accounted for another 2.2 percentage points. Shrinking youth population and reduced labor market participation had a greater effect than any direct employment adjustment driven by AI, the analysis found.

Even so, signs emerged that young workers are being pushed out of high-exposure occupations at a relatively faster pace. Using 2022 as the base year of 100, the employment index for workers under 30 in high-exposure occupations stood at 86.6 last year, compared with 92.5 for those in no-exposure occupations. The number of under-30 workers in high-exposure jobs fell from about 1.197 million in 2022 to about 1.037 million last year, while those in no-exposure occupations declined from 915,000 to 846,000 over the same period.

Researchers said the decline in youth employment in high-exposure occupations had been under way before the launch of ChatGPT, making it difficult to attribute the trend solely to generative AI. They warned, however, that if young people continue to enter high-exposure fields at a slower rate, it could undermine career development and skills accumulation, and called for continued monitoring.

AI adoption in industry remains at an early stage, the survey found. Of 2,297 businesses across 21 industries, 656 — or 28.6 percent — had adopted AI. The information and communications sector had the highest adoption rate at 62.9 percent, followed by precision instruments at 41.3 percent and food products at 39.3 percent. Transportation had the lowest rate at 11.4 percent.

Among businesses that had adopted AI, 75.6 percent used generative AI, 27.6 percent used automated systems and 13.3 percent used predictive AI. Companies primarily deployed AI not to reduce headcount but as a productivity tool — cutting down on repetitive tasks such as document drafting, information retrieval and customer service.

Workers reported feeling both the benefits and the anxieties of AI adoption. A survey of 305 people in IT development, visual and digital design, and general office roles found that AI use scored 3.40 out of 5 for improving work efficiency, 3.15 for problem-solving capability, and 3.14 for productivity. Scores for creativity and collaboration were lower, at 2.79 and 2.30, respectively.

Among respondents, 52.1 percent said they felt burdened by the need to constantly learn new AI tools, and 38.7 percent said they felt at risk of having their jobs replaced. Concerns were also raised that as AI takes over basic and entry-level tasks, the "skills ladder" — through which junior workers build expertise by handling foundational work — could weaken. Researchers recommended expanding on-the-job AI training and job-transition programs that allow workers to develop professional skills while using AI in real work settings.

Kwon Woo-hyun, head of the employment policy research division at the Korea Employment Information Service, said the ultimate impact of AI on jobs is more likely to be determined by social and policy choices than by the nature of the technology itself. "We need to refine a human-centered labor market development strategy that treats AI as a partner for expanding human potential," he said.


fact0514@heraldcorp.com