The introduction of artificial intelligence (AI) is expected to increase South Korea's gross domestic product (GDP) by up to 12.6%. Productivity is projected to rise by up to 3.2%. However, the effects of productivity enhancement are concentrated among large corporations and those with lengthy operating histories, and it has been analyzed that one-third of workers may be at high risk of being replaced by AI in this process.
On the 10th, the Bank of Korea published the BOK Issue Note titled 'AI and the Korean Economy,' which contains these findings. Team leader Oh Sam-il and Director Lee Soo-min from the Bank of Korea's Research Bureau's Employment Research Team participated in the drafting of the report, using the job-centered model proposed by Cazzaniga last year for analysis.
According to the job-centered model, the introduction of AI affects the economy through three pathways: labor replacement, labor supplementation, and overall productivity enhancement. The researchers assumed that these three pathways present complex interactions and analyzed the effects of productivity enhancement by dividing it into three scenarios: ▲when labor replacement and labor supplementation occur simultaneously (Scenario 1), ▲when labor replacement and productivity enhancement occur together (Scenario 2), and ▲when all three—labor replacement, labor supplementation, and productivity enhancement—occur (Scenario 3).
Simulation results indicated that the introduction of AI leads to productivity enhancement and economic growth in all scenarios. In Scenario 1, the introduction of AI increased total factor productivity (TFP) and GDP by 1.1% and 8.4%, respectively. In Scenario 2, TFP and GDP grew by 2.1% and 4.2%, respectively. In Scenario 3, TFP and GDP rose by 3.2% and 12.6%, respectively. These productivity and growth effects are expected to occur within ten years.
This is expected to substantially offset the GDP decrease caused by aging. The researchers noted that 'if there is no AI introduction, South Korea's GDP is expected to decline by 16.5% from 2023 to 2050 due to a decrease in labor supply, but in Scenario 3, the introduction of AI could significantly reduce this decline to 5.9%,' stating that it could substantially counterbalance the slowdown in growth caused by aging and decreased labor supply.
However, the effects of productivity enhancement appear to be concentrated among large corporations and those with lengthy operating histories. An analysis conducted using corporate activity survey data from 2017 to 2022 revealed that the productivity enhancement effects of AI are particularly pronounced in large corporations and those with lengthy operating histories. In terms of revenue, the effect of AI introduction was especially clear in companies with lengthy operating histories.
Additionally, the researchers analyzed the impact of AI on the labor market. This analysis utilized indicators of 'AI exposure' and 'AI supplementation' for different occupations. AI exposure refers to the degree to which the tasks performed by a particular occupation can be substituted by AI, while AI supplementation represents the extent to which the social and physical attributes of an occupation provide protection against the risk of job replacement by AI.
For example, judges and accounting or clerical positions are classified as high exposure and low supplementation jobs, as the introduction of AI is likely to lead to a reduction in job opportunities. Workers in agriculture, forestry, and fisheries, who must perform many physical tasks, have a low AI exposure and also low AI supplementation, making them difficult to replace with AI. In contrast, jobs like surgeons have high AI exposure, but the importance of decision-making suggests that humans will likely continue to perform them, resulting in high AI supplementation.
The analysis revealed that more than half of domestic workers, or 51%, are significantly impacted by AI introduction. A total of 24% of all workers belong to a group with 'high exposure, high supplementation' that could benefit from AI-induced productivity, while 27% are classified as 'high exposure, low supplementation' and are at risk of being replaced or experiencing reduced income due to AI. Notably, women, youth, and those with higher education and income levels tend to show an increase in both AI exposure and supplementation.
Meanwhile, South Korea appears to be relatively well-prepared for AI adoption compared to other advanced countries. Utilizing the AI readiness index developed by Cazzaniga last year, South Korea ranked 15th among 165 countries. By institutional sector, it ranked 3rd in 'innovation and economic integration,' 18th in 'regulation and ethics,' and 18th in 'digital infrastructure,' either exceeding or matching advanced country averages.
However, in terms of 'human capital utilization and labor market policy,' South Korea fell below the advanced country average. Consequently, the researchers advised that 'targeted policies' are necessary to enhance flexibility in the labor market through education and retraining programs, while strengthening social safety nets for vulnerable groups.