Adopting artificial intelligence (AI) to boost labor productivity could lift South Korea's average annual GDP growth by as much as 0.4 percentage point over the next two decades, according to a state-run think tank.
The Korea Institute of Public Finance (KIPF) laid out the AI adoption scenarios in a report on "the impact of demographic change on the real economy and fiscal spending," submitted to the Presidential Committee on Ageing Society and Population Policy.
Researchers modeled three cases: a 10% rise in labor productivity for younger workers aged 25–45 only (Scenario 1); a 7% gain across all age groups (Scenario 2); and a differentiated increase of 10% for younger workers and 5% for middle-aged and older workers (Scenario 3). The figures were adjusted for South Korea's employment rate from estimates by researchers including Stanford University professor Erik Brynjolfsson, who found that AI adoption raises average labor productivity by 14%. The model assumes productivity gains phase in over two steps, in 2030 and 2035.
The analysis showed that average annual GDP growth over the next 20 years would rise 0.3–0.4 percentage point from the baseline of 1.21% without AI, reaching 1.51% under Scenario 1, 1.58% under Scenario 2, and 1.63% under Scenario 3. "Higher labor productivity from AI adoption not only directly affects effective labor but also raises the marginal productivity of capital, boosting investment and the growth rate of total capital," the report said.
With South Korea's potential growth rate projected to fall to around 0% in the 2040s — and possibly turn negative in the worst case — the additional GDP uplift from AI is expected to serve as a key variable in preventing the economy from contracting.
Faster growth also translated into more stable pension finances. Under AI adoption, the depletion of the National Pension Fund would be pushed back from 2062 in the baseline to 2065 under Scenario 1 and to 2067 under Scenarios 2 and 3 — delays of three to five years. The delay would also reduce the social security interest burden in 2070 by 1.48 percentage points of GDP under Scenario 3. "Higher economic growth not only increases pension revenue, but the fact that such revenue gains are concentrated in the next 20 years, while the National Pension Fund still has ample reserves, also contributes to fiscal stability," the report said. Kim Pyeong-sik, an associate research fellow at the KIPF, said, "Even though we modeled the AI effect only as a rise in labor productivity rather than in total factor productivity (TFP), clear short-term growth momentum and fiscal buffering effects were confirmed."
Unlike AI adoption, raising the fertility rate or expanding immigration would have only limited short-term impact, the analysis found. For fertility, the GDP gap between the high- and low-estimate scenarios was just 0.06 percentage point in 2050, widening to 0.77 percentage point only by 2070. "A fertility rebound is absolutely necessary, but there is a long physical lag before newborns join the economically active population and translate into a genuine expansion of the tax base," Kim said. Expanded immigration, while growing the overall economy, had negligible effect on per-capita GDP and extended pension fund depletion by only two years.
Accordingly, the researchers proposed a time-phased "policy package" as a fundamental solution. The three-step approach calls for managing the growth in social insurance spending — including health insurance and long-term care insurance for the elderly — in the short term; reinforcing the growth and revenue base through AI adoption and expanded immigration in the medium term; and easing downward pressure from population decline through fertility-rebound policies in the long term.
Preparing for labor market shocks remains a challenge, however. The report noted that "if AI operates mainly by replacing labor rather than driving innovation across industries, the positive economic effects could be smaller." The Bank of Korea said in a report last February that 27% of South Korean workers are highly likely to be replaced by AI or see their incomes decline. Analysts say AI adoption must be accompanied by carefully calibrated workforce redeployment policies to prevent vulnerable workers from being pushed out of the labor market.
