KDI says AI may lift productivity but reduce jobs: how to read the 256,000 estimate
KDI’s report “Analysis of the Macroeconomic Impact of AI” has moved to the center of the labor-market debate. Reports said that AI diffusion could raise Korean productivity by up to 3.5% over the next 10 years, while also creating pressure that may reduce annual jobs by 256,000. The important point is not to reduce this to “AI will destroy all jobs.” Technology diffusion produces different outcomes depending on productivity gains, job transitions, education, and redeployment policies.
Key summary
- KDI’s report analyzed how AI diffusion may affect the macroeconomy and labor market.
- Multiple reports said productivity could rise by up to 3.5% over 10 years.
- The same reports introduced estimated annual job-reduction pressure of about 256,000 positions.
- The policy focus is not blocking AI, but designing job transition, reskilling, and protection for vulnerable workers.
What the numbers say and do not say
Higher productivity means more value can be produced in the same amount of time. For companies, that may mean lower costs and faster work. But the gains do not automatically flow to every worker. Jobs involving repetitive tasks, document organization, customer response, or basic analysis may feel pressure first, while planning, verification, and field judgment that use AI as a tool may become more important.
Confirmed facts
- KDI released the research report “Analysis of the Macroeconomic Impact of AI” on July 20, 2026.
- MBC News and Dong-A Ilbo reported both the potential productivity gain and the job-reduction estimate.
- Yonhap Infomax reported that AI could raise Korea’s productivity while creating labor-market adjustment pressure.
- Because these are forecast estimates, the report and coverage should be read with the assumption that outcomes may change depending on policy response and adoption speed.
How to read each indicator
| Indicator | Content | Reading point |
|---|---|---|
| Productivity | Possible increase of up to 3.5% over 10 years | Overall efficiency may rise, but distribution matters. |
| Jobs | Estimated annual pressure of 256,000 fewer jobs | Separate disappearing tasks from newly created tasks. |
| Company response | More automation and AI-assist tools | Cost savings must be paired with quality-control systems. |
| Policy task | Reskilling, job transition, and safety nets | Data that identifies affected occupations early is critical. |
What to watch next
- Watch whether the government and companies turn reskilling programs into concrete plans at the speed of AI adoption.
- Check where AI-driven cost savings are allocated: wages, employment, service quality, or margins.
- More detailed statistics are needed because young workers, mid-career workers, office jobs, and field jobs may be affected differently.
- AI regulation should be designed with balance between innovation and labor protection.
Search keywords
- KDI AI macroeconomic impact analysis
- AI productivity 3.5% jobs 256,000
- AI labor market reskilling
- artificial intelligence job outlook