TItle page 1

Contents 5

Abstract 4

About the authors 4

Acronyms 10

Introduction 12

1. Task-based approaches to automation in the literature 14

2. Tasks and occupations in the 6-digit system in Poland 15

3. Assessment of tasks' automation potential 17

3.1. Initial Algorithmic predictions 17

3.2. Selection of occupational tasks for human assessment 18

3.3. Survey design and recruitment 20

3.4. Survey design limitations 21

3.5. Survey of Task Automation Potential 22

3.5.1. Sample properties 22

3.5.2. Participants' Exposure to GenAI: Screening Questions 25

3.5.3. Scoring of task automation potential 26

4. Expert validation survey 30

5. Adjustment of survey sub-sample 33

6. Adjustment of all survey scores 37

7. Prediction of synthetic task-level scores for all ISCO-08 and 6-digit occupations 38

8. Adjusted Global Index of GenAI Exposure 40

9. Changes to occupational classifications 43

10. Revised global employment estimates 46

Conclusion 49

Annex 51

1. Exposure by 4-digit ISCO-08 occupation 51

2. Sampling Formula for section 4.2. 64

3. Survey Questionnaire 65

References 71

Acknowledgements 74

Tables 8

Table 1. Composition of occupational tasks in the Polish classification system and in ISCO-08 16

Table 2. Survey sample selection 19

Table 3. Core survey questions 20

Table 4. Sample of adjustments and justifications between the survey and expert scores (examples of largest upward and downward revisions) 34

Table 5. GenAI Exposure Gradients: Definition and Interpretation 41

Figures 7

Figure 1. Income- and population-based similarities (A) and access to the internet (B) across countries 15

Figure 2. Distribution of synthetic automation scores from 3 LLMs, by ISCO-08 1-digit 17

Figure 3. Age and sex distribution in the survey compared to Labour Force Survey (LFS) data in Poland (employed individuals) 23

Figure 4. Distribution of occupational groups in the survey, compared to LFS in Poland (employed individuals) and to the desired sample (Table 2) 24

Figure 5. Occupation and sex distribution among survey participants compared to LFS in Poland (employed individuals) 24

Figure 6. Frequency of use of GenAI, by 1-digit ISCO-08 25

Figure 7. Expectations of impact on the work area 26

Figure 8. Expectations of impact on individuals' current job - distribution of individual scores 26

Figure 9. Distribution of task-level scores by sex and occupational group (1-digit ISCO-08) 27

Figure 10. Scoring as a function of familiarity with GenAI and the scored task - distribution of individual scores 28

Figure 11. Schema of the scoring stages 30

Figure 12. Example of a dendogram used for the review of tasks' semantic clustering 31

Figure 13. Task-level scores from the survey and experts, compared to AI-arbitrated scores 33

Figure 14. Comparison of Adjusted Scores: GPT-4o vs Gemini by Occupational Group 34

Figure 15. Survey scores, expert evaluation and final adjustments for 2,861 tasks in the main survey 37

Figure 16. AI Exposure Gradients: ISCO-08 4-digit level occupations 40

Figure 17. AI Exposure: Comparison of 2025 with 2023, ISCO -08 4-digit level occupations 43

Figure 18. Changes to occupational exposure between 2023 and 2025 within Gradient 4 44

Figure 19. 4-digit level occupations (ISCO-08) with largest changes in mean scores between 2023 and 2024 45

Figure 20. Global estimates of occupations potentially exposed to GenAI (% of employment by sex) 46

Boxes 9

Text Box 1. Sample Question for Task Automation Assessment 21

Text Box 2. Introductory prompt for survey respondents 22

Text Box 3. The conceptual structure of the LLM prompt (Python code for GPT-4o) 38

Annex Tables 8

Table A1. ISCO-08 occupations by exposure gradient 51