生成式人工智能与就业：对工作数量和质量潜在影响的全球分析(英文 55页)

Generative AI and Jobs: A global analysis of potential effects on job quantity and quality

Introduction 07

1 Methods and Data 10

1.1. ISCO data on occupations and tasks 11

1.2. Prompt design and sequence 12

2 Assessment of the Predictions, Robustness Tests and the Bounds for Analysis 17

3 Results 20

3.1. Automation vs augmentation: distribution of scores across tasks and occupations 24

4 Exposed occupations as a share of employment: global and income-based estimates 30

4.1. Augmentation vs Automation: ILO microdata 30

4.2. Augmentation vs Automation: global estimate 32

4.3. The big unknown 36

5 Managing the transition: Policies to address automation, augmentation and the growing digital divide 38

5.1 Mitigating the negative effects of automation 38

5.2 Ensuring job quality under augmentation 39

5.3 Addressing the digital divide 40

Conclusion 43

Appendix 1. Countries with missing ISCO-08 4-digit data: estimation procedure 45

References 47

Acknowledgements and use of GPT 51

List of Figures

Figure 1. Mean automation scores by occupation, based on ISCO and GPT tasks 21

Figure 2. Tasks with medium and high GPT-exposure, by occupational category (ISCO 1-digit) 24

Figure 3. Box plot of task-level scores by ISCO 4d, grouped by ISCO 1d 25

Figure 4. Augmentation vs automation potential at occupational level 27

Figure 5. Occupations with high automation potential 28

Figure 6. Occupations with high augmentation potential 29

Figure 7a. Automation vs augmentation potential: shares of total employment, microdata for 59 countries 30

Figure 7b. Automation vs augmentation potential: shares of total employment in each sex  (ILO microdata) 31

Figure 8. Country coverage based on the level of digits in ISCO-08 (ILO data) 33

Figure 9a. Global estimates: jobs with augmentation and automation potential as share of total employment 34

Figure 9b. Automation vs augmentation potential: shares of total employment for each sex (global estimate) 35

Figure 10. Occupations with high automation potential, by ISCO 4-digit and income group 36

Figure 11a. The “Big Unknown”: occupations between augmentation and automation potential 37

Figure 11b. The “Big Unknown”: share of total employment, by income group (global estimate) 37

Figure 11. Share of population not using the internet 41

Figure 12. A classic growth path: income and occupational diversification 42

生成式人工智能和就业：对就业数量和质量的潜在影响的全球分析

简介07

1 方法和数据 10

1.1. ISCO 职业和任务数据 11

1.2. 提示设计和顺序 12

2 预测评估、稳健性测试和分析界限 17

3 结果 20

3.1. 自动化与增强：跨任务和职业的分数分布 24

4 暴露职业占就业的比例：全球和基于收入的估计 30

4.1. 增强与自动化：ILO 微观数据 30

4.2. 增强与自动化：全球估计 32

4.3. 最大的未知数 36

5 管理转型：解决自动化、增强的政策以及日益扩大的数字鸿沟 38

5.1 减轻自动化的负面影响 38

5.2 确保增强下的工作质量 39

5.3 解决数字鸿沟 40

结论 43

附录 1. 缺少 ISCO-08 4 位数据的国家：估计程序 45

参考文献 47

GPT 51 的致谢和使用

图列表

图 1. 基于 ISCO 和 GPT 任务的职业平均自动化得分 21

图 2. 按职业类别划分的具有中度和高度 GPT 暴露的任务（ISCO 1 位数字）24

图 3. ISCO 4d 任务级别得分的箱线图，按 ISCO 1d 25 分组

图 4. 职业层面的增强与自动化潜力 27

图 5. 自动化潜力较高的职业 28

图 6. 具有高增强潜力的职业 29

图 7a。 自动化与增强潜力：总就业份额，59 个国家的微观数据 30

图 7b。 自动化与增强潜力：各性别占总就业人数的比例（国际劳工组织微观数据）31

图 8. 基于 ISCO-08 中数字级别的国家覆盖范围（国际劳工组织数据）33

图 9a。 全球估计：具有增强和自动化潜力的工作占总就业人数的比例 34

图 9b。 自动化与增强潜力：每种性别占总就业人数的比例（全球估计）35

图 10. 具有高自动化潜力的职业，按 ISCO 4 位数和收入组 36 划分

图 11a。 “大未知”：介于增强和自动化潜力之间的职业 37

图 11b。 “大未知数”：按收入群体划分的总就业份额（全球估计）37

图 11. 不使用互联网的人口比例 41

图 12. 经典增长路径：收入和职业多元化 42

×