목차
Title page 1
Contents 6
Foreword 4
Acknowledgements 5
Abbreviations and acronyms 12
Executive summary 13
1. AI use in the Japanese workplace 15
In Brief 16
1.1. Introduction 18
1.2. Who are the workers most likely to use AI at work? 20
1.2.1. AI users and non-AI users 20
1.2.2. Characteristics of companies that use AI 24
1.2.3. Characteristics of employees who use AI 29
1.3. What are the barriers to the adoption of AI in the workplace? 39
1.3.1. Perceptions of companies 39
1.3.2. Why do people not use GEAI? 41
Annex 1.A. AI use in the Japanese workplace: Additional figures 44
References 47
Notes 50
2. Reaping the benefits of AI for performance at work and job quality 52
In Brief 53
2.1. Previous studies on impact of AI on the performance at work and job quality 55
2.1.1. Quality of the work environment 55
2.1.2. Impact on wages 57
2.1.3. Equalisation of performance within occupations 58
2.2. The impact of AI on performance at work and job quality in the Japanese labour market 59
2.2.1. Distinctive features of the impact of AI on performance at work and job quality in Japan 59
2.2.2. How the impact of AI on wages varies by worker characteristics 77
2.2.3. The impact of AI on work processes by worker characteristics 83
2.3. Initiatives to enhance outcomes related to AI on performance at work and job quality 87
2.3.1. Company training, financial support for training, self-learning 87
2.3.2. Worker consultation on the use of new technology 93
2.3.3. Internal rules or guidelines for the appropriate use of GEAI 95
2.3.4. Trust in own company to only use AI that is safe and trustworthy 97
Annex 2.A. Reaping the benefits of AI for performance at work and job quality: Additional figures 99
References 109
Notes 112
3. Preparing for the impact of AI on job quantity and skills needs 114
In Brief 115
3.1. The impact of AI on Job quantity and skills needs: A review of the literature 117
3.1.1. The impact of AI on Job quantity 117
3.1.2. The impact of AI on skills 118
3.2. New evidence on the impact of AI on job quantity and skills needs in the Japanese labour market 119
3.2.1. The expected impact of AI on job quantity 119
3.2.2. The impact of AI on the volume of work tasks 129
3.2.3. Prospective impact of AI on the skill needs 132
3.3. How might training and worker consultation influence job quantity and skill requirements? 136
3.3.1. Training and AI-Driven impacts on job quantity 137
3.3.2. Worker consultation and AI-Driven impacts on job quantity 137
3.3.3. Internal rules or guideline and GEAI-Driven impacts on job quantity 137
3.3.4. Trust in employers and AI-Driven impacts on job quantity 139
Annex 3.A. Preparing for the impact of AI on job quantity and skills needs: Additional figures 141
References 144
Notes 146
4. Maximising the benefits of AI while ensuring the safety and trustworthiness of AI technologies in the workplace 147
In Brief 148
4.1. Challenges related to company-provided training and self-learning for working collaboratively with AI 150
4.1.1. Challenges related to company-provided training and financial support 150
4.1.2. Challenges concerning the support for workers' voluntary reskilling and upskilling 154
4.2. Challenges related to worker consultation on the use of new technologies in the workplace 160
4.2.1. Gender and worker consultation 162
4.2.2. Age group and worker consultation 162
4.2.3. Worker consultation by worker characteristics 163
4.2.4. Reasons for lack of labour-management communication 165
4.3. Challenges related to establishing internal rules or guidelines to ensure employees can use GEAI appropriately in their work 166
4.3.1. Understanding and compliance to guidelines 169
4.4. Challenges related to building employee trust in the safety and reliability of AI technologies used in the workplace 171
4.4.1. Bias and discrimination 174
4.4.2. Protecting worker health and safety 175
4.4.3. Data protection and privacy 176
4.4.4. Human agency and oversight 178
4.4.5. Explainability 180
4.4.6. Accountability 181
4.5. Recent regulatory and policy developments in Japan regarding the safety, reliability, and transparency of AI technologies 182
4.5.1. The evolution of AI-related legislation in Japan 182
4.5.2. The ongoing discussions regarding the APPI in Japan 183
4.5.3. The MHLW's efforts so far and recommendations for future action 184
4.5.4. Enhancing public employment services through the use of AI 184
Annex 4.A. Maximising the benefits of AI while ensuring the safety and trustworthiness of AI technologies in the workplace: Additional figures 186
Annex 4.B. Japan's new AI Act 196
References 198
Notes 200
Figures 7
Figure 1.1. Compared to other countries, Japanese AI adopters are more likely to report being managed by AI or having no interaction... 21
Figure 1.2. The proportion of Japanese employees using AI at work is the lowest among countries surveyed 22
Figure 1.3. Many Japanese employees expect AI use in workplaces to increase in the future 23
Figure 1.4. While the proportion of Japanese employees using AI at work in the information and communications sector exceeds 20%,... 24
Figure 1.5. Japanese AI adopters across many industries report that the use of AI in their companies has expanded since May 2022 25
Figure 1.6. Japanese employees in larger companies are more likely to report using AI, similar to those in other countries 26
Figure 1.7. Japanese AI users in companies experiencing a mild labour shortage are more likely to report using AI 28
Figure 1.8. There are regional disparities in Japan, with employees in the South Kanto and Kinki regions being more likely to say they are using AI 29
Figure 1.9. Japanese AI users are more likely to be young and middle-aged, male and more educated than non-users, similar to those in other countries 30
Figure 1.10. Japanese AI users (particularly GEAI users) are more likely to be regular employees than non-users or non-adopters 31
Figure 1.11. Japanese employees with disabilities or those engaging in childcare or long-term care are more likely to use AI 32
Figure 1.12. Japanese employees who frequently handle data analysis or tasks required creativity or tasks in hazardous place or managing... 34
Figure 1.13. Japanese Managers, Professionals and Technicians and associate professionals are more likely to use AI 37
Figure 1.14. Japanese Managers, Professionals, and Technicians and associate professionals are most likely to report the use of AI in their companies... 38
Figure 1.15. Compared to U.S. companies, Japanese companies are more likely to report that a shortage of AI-related talent is a barrier to AI adoption,... 40
Figure 1.16. Compared to U.S. companies, Japanese companies are more likely to report that shortage of employees who can promote AI adoption... 41
Figure 1.17. The most common reasons Japanese people give for not using GEAI are "I don't know how to use it' and "I don't need it in my life' 42
Figure 1.18. The most common reasons Japanese full-time employees, employers, or executives give for not using GEAI are "Concerns about... 43
Figure 2.1. Although Japanese AI users report that AI improves their performance and working conditions, the effects are more moderate... 60
Figure 2.2. Japanese male AI users are less likely than their counterparts in other countries to report that AI improves their performance... 62
Figure 2.3. Japanese AI users with university degree are less likely than their counterparts in other countries to report that AI improves... 63
Figure 2.4. Middle-aged and older Japanese AI users are less likely than their counterparts in other countries to report that AI improves... 64
Figure 2.5. Japanese AI users in non-regular employment are less likely to report that AI improves their performance and working conditions 65
Figure 2.6. Japanese AI users in SMEs are more likely to report that AI improves their performance and working conditions 66
Figure 2.7. SMEs in all countries see enhanced performance as the main benefit of generative AI 67
Figure 2.8. SMEs in Japan are most likely to say that generative AI helps compensate for labour and skill shortages 68
Figure 2.9. Japanese AI users with disabilities or those engaging in childcare and/or long-term care are more likely to report that AI improves... 70
Figure 2.10. The proportion of AI users reporting improvements in job quality, as well as the nature of these improvements, tends to vary by occupation 71
Figure 2.11. Although there are regional differences, Japanese AI users in rural areas as well as urban areas report that AI improves job performance... 72
Figure 2.12. Japanese AI users report that AI improves the work environment in various ways 74
Figure 2.13. Japanese AI users who frequently handle management tasks or tasks in hazardous places are more likely to report a higher improvement... 75
Figure 2.14. Japanese AI users are more likely to report that AI improves their performance and working conditions if they experience changes... 77
Figure 2.15. As in other countries, more employees in Japan expect their wages to decrease rather than increase due to AI in the next 10 years,... 78
Figure 2.16. GEAI users are more likely to report wage increases rather than decreases due to AI 79
Figure 2.17. AI users who are male, younger, regular employees, caregivers or have a disability are more likely to report wage increases rather... 80
Figure 2.18. Japanese AI users in SMEs are more likely to wage increases rather than decreases after using AI, compared to those in large companies 81
Figure 2.19. AI users of plant and machine operators, and assemblers are most likely to report wage increases rather than decreases after using AI 82
Figure 2.20. Japanese AI users, both in urban and rural areas, are more likely to report wage increases rather than decreases after using AI 83
Figure 2.21. While Japanese AI users report that AI has increased their pace and their control over the sequence in which they perform their tasks,... 84
Figure 2.22. While Japanese AI users report that it assists with decision making, the impact is milder than other countries 85
Figure 2.23. Japanese AI users also agree that AI helps them make better and faster decisions 86
Figure 2.24. The gap in reported improvements in job quality between Japanese AI users who received company-provided training and those who did... 88
Figure 2.25. Japanese companies are more likely to provide in-house training seminars 89
Figure 2.26. Japanese AI users report that OJT or seminars at external organisations have the potential to be even more likely to improve their job performance 90
Figure 2.27. Japanese AI users who have received company training and engaged self-learning are even more likely to report positive outcomes... 91
Figure 2.28. Although Japanese AI users who have received company training and engaged in self-learning report improvements in the work environment,... 92
Figure 2.29. Japanese AI users who have received company training and engaged in self-learning are more likely to report wage increases rather... 93
Figure 2.30. Japanese AI users report that consultations help alleviate worries, enhance the efficiency of technology introduction, and clarify required skills 94
Figure 2.31. The differences in reported improvements in performance and working conditions between Japanese AI users who receive consultation... 95
Figure 2.32. Japanese GEAI users whose companies have established internal rules or guidelines are more likely to report positive outcomes of AI... 96
Figure 2.33. Japanese AI users trusting own company to only use AI that is safe and trustworthy are even more likely to report positive outcomes of AI... 98
Figure 3.1. Japanese AI adopters are less likely than those in other countries to know of individuals in their company who have lost jobs due to AI 120
Figure 3.2. Japanese AI users are the most worried about losing their job due to AI in the next 10 years 121
Figure 3.3. Japanese AI users are more likely to believe AI will result in job creation than job loss, whereas other employees hold the opposite view 122
Figure 3.4. The balance between worries about job loss and expectations of job creation from AI varies depending on the characteristics of Japanese AI users 123
Figure 3.5. AI users in Japan, regardless of company size, tend to report more expectations for job creation than worries about job loss 124
Figure 3.6. Japanese AI users in Managers and Professionals are more likely to report higher expectations for job creation than worries about job loss 126
Figure 3.7. The balance between the proportion of AI users reporting worries about job loss and those expressing expectations for job creation... 129
Figure 3.8. While Japanese AI users also report task automation and creation due to AI, the impact is milder than other countries 130
Figure 3.9. Particularly, Japan may be lagging behind other surveyed countries in automating repetitive and complex tasks 132
Figure 3.10. Japanese AI users, similar to their counterparts in other countries, tend to agree more with the idea that AI complements their skills... 133
Figure 3.11. 89.4% of AI users in Japan reported that they expect the skills and abilities required in their current job to change due to AI over the next ten years 134
Figure 3.12. Japanese AI users in Managers, Professionals, Technicians and associate professionals are more likely to report that AI will change the skills... 135
Figure 3.13. Japanese AI adopters anticipate that various skills will grow in importance over the next decade, particularly the ability to identify... 136
Figure 3.14. Japanese AI users who have participated in both company-provided training and self-directed learning are more likely to anticipate... 138
Figure 3.15. Japanese AI users whose employers consult them regarding the use of new technologies in the workplace are more likely to anticipate... 139
Figure 4.1. Although around 30% of Japanese AI users report that their company has provided or funded training so that they can work with AI,... 151
Figure 4.2. AI users in Japan are less likely to receive company training to work with AI 152
Figure 4.3. The difference between Japan and the average of 7 OECD countries in the proportion of AI users who received company training... 153
Figure 4.4. The difference between Japan and the average of 7 OECD countries in the proportion of AI users who received company training... 154
Figure 4.5. 35% of AI users in Japan report that they engaged in reskilling/upskilling to work with AI 155
Figure 4.6. 55% of Japanese AI users who report that they engaged in reskilling/upskilling in 2023 had part of their training course fees subsidised 156
Figure 4.7. Older and non-regular Japanese employees are less likely to report using the Educational Training Benefits for reskilling or upskilling to work with AI 157
Figure 4.8. There are regional differences among Japanese AI users in whether they have access to resources for learning how to work with AI 159
Figure 4.9. Japanese AI users in companies facing labour shortage are less likely to report having the resources to learn to work with AI 160
Figure 4.10. Japanese employees and AI users are less likely than those in other countries to report that their employers consult workers... 161
Figure 4.11. The difference between Japan and the average of 7 OECD countries in the proportion of AI users whose employers consult... 162
Figure 4.12. The difference between Japan and the average of 7 OECD countries in the proportion of AI users whose employers consult... 163
Figure 4.13. AI users who middle-aged and older, non-regular employees are less likely to report that their employers consult them regarding... 164
Figure 4.14. Many Japanese employers who didn't consult employees on introducing new technology in the workplace saw it as an insignificant... 166
Figure 4.15. 34.8% of generative AI users in Japan report that internal rules or guidelines have been established to support working appropriately with AI 168
Figure 4.16. SMEs using generative AI in Germany are the most likely to have guidelines in place 169
Figure 4.17. A certain percentage of GEAI users report that they don't understand or comply with company rules or guidelines to work appropriately with GEAI 170
Figure 4.18. Japanese employees are more likely to prioritise improving the safety, reliability, and transparency of AI technology as initiatives... 171
Figure 4.19. Japanese AI users are less likely than those in other countries to explicitly report trusting their company to use only safe and trustworthy AI 172
Boxes 33
Box 1.1. Using AI to support people with disability in the labour market 33
Box 1.2. AI technology used for tasks in hazardous places 35
Box 2.1. The potential of generative AI to support SMEs 67
Box 3.1. Japanese employees' expectations of AI's impact on job quantity over the next ten years (ISCO digit 3) 127
Box 4.1. AI-related training participation among SMEs employees 152
Box 4.2. Guidelines for staff concerning the use of GEAI among SMEs 168
Annex Tables 10
Annex Table 1.A.1. The relationship between AI usage rates and worker characteristics 45
Annex Table 2.A.1. Marginal effects of non-regular employment on job performance (Generalised Ordered Logit Model) 99
Annex Table 2.A.2. Marginal effects of company size on job performance (Generalised Ordered Logit Model) 100
Annex Table 2.A.3. Marginal effects of disability on job performance (Generalised Ordered Logit Model) 100
Annex Table 2.A.4. Marginal effects of caregiving responsibility on job performance (Generalised Ordered Logit Model) 101
Annex Table 2.A.5. Marginal effects of occupation on job performance (Generalised Ordered Logit Model) 102
Annex Table 2.A.6. Marginal effects of AI user attributes on wage (Generalised Ordered Logit Model) 104
Annex Table 2.A.7. Marginal effects of disability on wage (Generalised Ordered Logit Model) 104
Annex Table 2.A.8. Marginal effects of caregiving responsibility on wage (Generalised Ordered Logit Model) 105
Annex Table 2.A.9. Marginal effects of occupation on wage (Generalised Ordered Logit Model) 106
Annex Table 2.A.10. Marginal effects of training on job performance (Generalised Ordered Logit Model) 107
Annex Table 2.A.11. Marginal effects of worker consultation on job performance (Generalised Ordered Logit Model) 108
Annex Table 3.A.1. Japanese employees' expectations of AI's impact on job quantity over the next ten years (ISCO digit 3) 141
Annex Table 4.A.1. Marginal effects of residential area on the availability of resources to learn to work with AI (Generalised Ordered Logit Model) 186
Annex Table 4.A.2. Marginal effects of manpower status on the availability of resources to learn to work with AI (Generalised Ordered Logit Model) 187
Annex Table 4.A.3. Marginal effects of Japanese worker attributes on the likelihood of trusting or not trusting that their employer uses only safe... 189
Annex Table 4.A.4. Marginal effects of Japanese AI users attributes on the likelihood of trusting or not trusting that their employer uses only safe... 192
Annex Figures 10
Annex Figure 1.A.1. Many Japanese companies also expect AI-based digital technology to help address labour shortages 44
Annex Figure 1.A.2. Many Japanese SMEs cite the lack of the right skills among employees as a barrier to the use of GEAI in the workplace 46
Annex Figure 1.A.3. In every age group, a large number of Japanese employees use GEAI for drafting documents or texts 46
Annex Figure 2.A.1. Japanese GEAI users are more likely to report that AI improves their performance and working conditions compared... 99
Annex Figure 2.A.2. Japanese AI users in manufacturing are the second most likely to expect wages in their sector to increase due to AI in next 10 years 103
Annex Figure 2.A.3. The average proportion of AI users reporting "No effect" across six tasks varies depending on the occupation 107
Annex Figure 3.A.1. Japanese AI users trusting own company to only use AI that is safe and trustworthy are more likely to anticipate both job... 143
Annex Figure 4.A.1. Among the other countries, Japanese AI users are the least likely to report that their company has provided or funded training... 186
Annex Figure 4.A.2. Among surveyed countries, Japanese AI users are the least likely to report that their employers consult workers or worker... 188
Annex Figure 4.A.3. Japanese AI non-users are less likely than those in other countries to explicitly report trusting their company to use only... 194
Annex Figure 4.A.4. The smaller the company, the less likely Japanese employees are to report trusting their company to use only safe and trustworthy AI 195
