Korean firms using AI employed about 2.9% more workers, state research institute finds
KRIVET's analysis of 2017-2024 corporate survey microdata found overall employment about 2.9% higher at companies after they began using artificial intelligence, with the largest gains where AI was used to develop products and services.
Companies in Korea employed roughly 2.9% more workers after they began using artificial intelligence than before, according to an analysis published by the Korea Research Institute for Vocational Education and Training, the government-funded institute headed by Ko Hye-won.
The findings appeared in the second issue analysis of a 2026 special edition of the institute’s periodical THE HRD REVIEW, released on July 31, 2026 and distributed on Aug. 6 from Sejong. The institute said the study drew on microdata from the Survey of Business Activities compiled by the National Data Office covering 2017 through 2024, limited to incorporated companies with at least 50 regular employees and capital of at least 300 million won. A firm was counted as using AI in any year in which its reported stage of AI use was something other than “not applicable.” The sample comprised 106,389 firm-year observations, with roughly 12,000 to 15,000 companies observed annually.
Estimates were produced with firm and year fixed effects and controls for asset size, according to the institute. On that basis, total employment was about 2.9% higher and regular employment about 2.4% higher, both statistically significant. At the same time, the share of regular workers in the workforce fell by about 0.46 percentage point. The institute said estimates for temporary and daily employment carried enough uncertainty that the results cannot be read as a proportional expansion of stable jobs.
KRIVET also cautioned against reading the raw gap between AI-using and non-using firms as an effect of the technology, saying the large difference in simple comparisons is substantially attributable to the tendency for larger companies with more employees to adopt AI in the first place.
The purpose of adoption mattered. Where AI was used for product and service development, employment rose about 4.0%, while the estimate for use in production processes was 1.5% and not statistically significant. The institute noted that 59.3% of companies using AI in 2024 reported product and service development as their main purpose. Firms that used AI continuously showed a 3.8% increase and new adopters 2.9%, though KRIVET said the statistical significance of the difference between those two point estimates was not separately confirmed.
By industry, employment was 12.0% higher in transportation and warehousing, 4.3% higher in professional, scientific and technical services and 4.1% higher in information and communications, while the 1.4% estimate for manufacturing was not statistically significant. An event study built around each firm’s first year of AI use showed employment gains widening after adoption, but the institute said pre-adoption estimates and wide confidence intervals mean the results should not be given a causal reading. It also said the analysis does not capture net employment effects across an entire industry, changes by occupation or task, or the intensity of AI use and scale of investment.
Jeong Ji-woon, a senior research fellow at KRIVET and head of its data analysis and performance dissemination center, said employment outcomes can differ depending on the purpose of AI use and the industry involved. He called for strengthening incumbent workers’ capacity to create value with AI, alongside reskilling and upskilling for jobs affected by automation of production processes and an employment safety net.
KRIVET was established in 1997 under the Act on the Establishment, Operation and Fostering of Government-Funded Research Institutes as a research body under the Office of the Prime Minister, with a mandate covering vocational education and training policy, qualification systems and the development and dissemination of training programs.