The Japan Institute for Labour Policy and Training (JILPT) analyzes the impact of generative AI on global employment based on the International Labour Organization's (ILO) May 2025 research brief 'AI and jobs: A 2025 update'.
Key Points
1. Comprehensive Occupational AI Impact Assessment Methodology
- Analysis targets 436 occupations based on International Standard Classification of Occupations (ISCO-08) and 29,753 tasks from Polish government occupational classification
- AI automation potential scored 0-1 for each task, with occupational impact classified into 6 levels: Stages 1-4 and 'Minimal Impact'/'No Impact'
- Evaluation method combining human and AI assessment, with comprehensive scoring through expert verification and AI learning
2. Concentrated Impact on Clerical Work and Adjustment to Realistic Assessment
- Most occupations with highest automation potential (Stage 4) are clerical (data entry, accounting, statistical/financial clerks, etc.)
- Overall average score decreased from 0.30 in 2023 to 0.29 in 2025, with smaller variance and more realistic assessment
- Tasks like meeting minutes and appointment scheduling decreased from 0.9 in 2023 to maximum 0.76 in 2025
- Shift from theoretical possibilities to assessment based on actual usage experience, confirming indispensability of human involvement
3. 24% of Global Employment Under Generative AI Impact
- 24% of global employment in 112 occupations (Stages 1-4) affected by generative AI
- Gender gap exists with 21% for men vs 28% for women showing higher impact rate for women
- High automation risk in clerical occupations where women predominantly work is main factor
- Pronounced gap in Stages 3-4: men 3.1%/2.4% vs women 5.7%/4.7%
4. Manifestation of Income Level and Regional Disparities
- Only 11% of employment affected in low-income countries while 34% in high-income countries
- Concentration of high-impact occupations (clerical, financial, customer service) in high-income countries is key factor
- Actual complete substitution possibility limited due to technological infrastructure and implementation cost constraints
- Structure where regional digital divide determines degree of employment impact
5. Emergence of New Technology-Responsive Occupations and Duality
- Score increases for digital-related occupations like web/multimedia developers, database designers/administrators
- Technological advances from 2023 improved image/audio processing and autonomous work progression capabilities
- Possibility of new occupation creation with generative AI development confirmed, showing duality of technological progress
- AI impact not directly leading to reduced worker roles but possibility of creating new occupational fields
The article concludes that while generative AI may affect one quarter of global employment, complete substitution is limited, with concentrated impacts on clerical workers and female workers, making response to regional disparities an important policy challenge.