AI spread to cut over 250,000 jobs a year in a decade, KDI warns; firms see 20% productivity gain

As AI technology spreads rapidly, more than 250,000 jobs could be lost annually within a decade, with mid- to high-skilled occupations such as professionals and office workers bearing the brunt, a state-run research institute has projected.

Nam Chang-woo, a senior research fellow at the Korea Development Institute, released a report titled "Analysis of AI's Macroeconomic Impact" on Wednesday.

The report estimates that automation driven by AI will eliminate roughly 256,000 jobs per year within 10 years — equivalent to about 2.1 percent of total employed workers last year. Professionals, office workers and sales staff in mid- to high-skilled categories are expected to be particularly affected.

The estimate was derived by applying historical production-automation patterns to tasks identified as susceptible to AI automation, to calculate the number of jobs realistically at risk of replacement.

The report also raised the possibility that wage inequality could narrow modestly. As wage growth slows in high-paying occupations, gaps between job categories may ease, and KDI projected that wage inequality as measured by the Gini coefficient would either improve slightly or hold at current levels.

The share of economic activity exposed to AI automation is also set to expand sharply. Currently standing at 1.4 percent by employment and 1.8 percent by sales, those figures are projected to climb to 8.1 percent and 10.5 percent, respectively, within a decade.

On the productivity side, the outlook is positive. The spread of generative AI is expected to raise the Korean economy's total factor productivity by 1.5 to 3.5 percent over the next 10 years — equivalent to an annual gain of 0.15 to 0.35 percentage points. The report attributed this to a multiplier effect, whereby productivity gains in AI-intensive industries ripple across the broader economy through inter-industry linkages.

An analysis of corporate financial data and AI adoption rates found that companies using AI saw revenue per full-time employee rise by an average of about 20 percent.

KDI recommended that the government act to maximize AI's economic benefits while stepping up efforts to cushion employment shocks and distributional pressures that may arise as AI spreads, and called on authorities to encourage complementary private-sector investment alongside those measures.

In the near term, the institute said the government should expand job redesign and retraining programs targeting workers in high-AI-exposure occupations and improve access to AI tools for small and medium-sized enterprises through public cloud infrastructure and shared GPU centers. It also called for the creation of a data-sharing platform.

Over the medium to long term, KDI said AI solutions that complement human labor should be more widely deployed in public services, and that industry-specific AI governance frameworks should be established to lay the groundwork for sustainable growth.

"There is a need to operate a labor-management-government consultative body focused on high-exposure professional occupations to discuss job transitions and wage structures, and to strengthen the social safety net — including closing gaps in unemployment benefits caused by the rise in non-standard work driven by AI," Nam said.