It appears that your browser has JavaScript disabled. Please enable JavaScript to view this page as intended.

JILPT Research Report No.228
AI Technology Implemented in the Workplaces and Its Impact on Workers: Japan's Position Based on an International Comparative Study with the OECD

March 28, 2024

Summary

Research Objective

This study aims to clarify the impact of AI technology on workers in the financial and manufacturing sectors in eight OECD countries. What patterns of the impact are observed, where does Japan’s case fit in, and what are the similarities and differences among the eight countries? The study explores what determines the divergence of the patterns observed across eight countries.

Research Method

Interview survey, Literature review

Key Findings

  1. Number of case studies by sector and by country, and number of interviews by stakeholder and by country

    The number of firms surveyed by sector and country totaled 96 case studies in the eight countries (Table 1). Due to difficulties in securing the number of firms for this survey, the Energy and Logistics sectors were added in some of these eight countries.

    Table 1. Number of case studies by sector and by country

    Finance Manufacturing Energy Logistics Total
    Austria 6 10 2 18
    Canada 6 7 13
    France 3 3 1 7
    Germany 3 6 1 10
    Ireland 4 8 3 1 16
    Japan 4 5 9
    UK 5 4 9
    US 7 7 14
    Total 38 50 7 1 96
    % of total 40% 52% 7% 1% 100%

    Source: Milanez (2023, 26).
    Original note: In Germany, Ireland and the UK, firms provided the research teams with cases (more than one) for study.

    In terms of the number of interviewees by stakeholder type and by country, there were slightly more interviews with management, HR personnel, and project managers responsible for implementing AI technologies. There were slightly fewer interviews with employees, such as general workers and representatives of trade unions or works councils (Table 2).

    Table 2. Number of interviews by stakeholder type and by country

    Workers Worker representatives Management HR AI implementation AI developers Other Total
    Austria 7 5 17 4 10 3 4 50
    Canada 4 3 18 4 4 5 38
    France 8 13 3 15 39
    Germany 12 7 7 1 9 4 40
    Ireland 2 3 15 9 4 33
    Japan 9 8 9 8 8 8 50
    UK 8 4 9 5 14 7 47
    US 7 2 14 1 13 9 46
    Total 57 32 102 22 82 39 9 343
    % of total 17% 9% 30% 6% 24% 11% 3% 100%

    Source: Milanez (2023, 27).
    Original Note: The category "other" consists of case study interviews held with IT personnel, IT managers, ethics researchers, purchasing assistants and data scientists.

  2. Changing the task composition of jobs

    The AI technology introduced four types of task changes. In each type of change, commonalities were found in cases in Japan and other countries.

    The first pattern observed was a complementary task change. Although the task itself remained the same as before, the technology induced the work efficiency and labor productivity improvement.

    Second, a full task automation was allowed by an AI technology. This means that of the multiple tasks a worker was responsible for, only specific tasks ware fully automated. It does not mean that the multiple tasks comprising worker's duty was displaced.

    Third, AI technologies partially automated the specific tasks.

    Fourth, AI technologies created new tasks such as data analysis and improvement in the accuracy of AI technology itself.

    These task changes were suggested to be related to the company's organizational policy on the AI technologies, the method the quality control is conducted at the manufacturing worksite, and the way tasks are reorganized.

  3. AI technologies change skill requirement

    Four patterns of skill change were identified. First, new skills such as sharpened analytical skills in data science were required. Japanese cases shared this same characteristic with other countries.

    Second, in cases where AI technology fully or partially automated specific tasks, for example, by increasing the emphasis on tasks requiring advanced skills that were traditionally performed by workers, proportion of these skills required have also increased. In this respect, there were similarities between the cases in Japan and other countries.

    Third, when AI technology fully or partially automated certain tasks, skill decreases occurred as low skill tasks were allocated to workers. While skill decrease was observed in the cases of other countries, no skill decrease was observed in the Japanese case.

    Fourth, AI technology did not impact or change the tasks performed by workers, and thus did not change the skills required of workers. This has been also observed in Japan as well as in other countries.

    It was suggested that changes in worker's skill requirements were related to the way tasks were reorganized.

  4. Changes in employment levels

    Three patterns of employment change were observed. In each pattern, there were similarities between Japan and other countries.

    First, AI technology had no impact on job quantities of workers and employment level have remained steady.

    Second, job losses have occurred in part. When the implementation of AI technologies reduced the job quantities of workers, business organizations responded either by reallocating workers within the firm or through non-replacing natural attrition, such as retirements. In the latter case, employment declined. The studies contained no instances of dismissals in any cases of any countries.

    Third, employment related to the further development of AI technologies was growing. It suggests that changes in employment after the introduction of AI technologies are related to the policy of using AI technologies and the way tasks are reorganized.

  5. The impact of AI on wages

    Three patterns of wage change were observed.

    First, wages for workers remained unchanged. The cases of Japan and other countries had this in common.

    Second, there were wage increases due to the more demanding tasks performed by workers and the skills required. While no wage increase was observed in the Japanese case study, it was observed in the cases from other countries. Differences were found between the Japanese and other country cases.

    Third, a decrease in the skills required for the worker's task has resulted in lower wages in some cases. In the Japanese case, there was no wage decrease, but this was the case in the other countries. Again, there was a difference between the Japanese case and the other countries.

    The results suggest that changes in wages after the introduction of AI technology are related to the wage system and collective bargaining agreements.

  6. Labor and management's response to AI technology

    First, the responses of labor and management on AI technology indicated that, representatives of trade unions and works councils viewed AI technology positively from the perspective of improving operational efficiency and the working environment. On the other hand, they also expressed concerns about future employment, increased scrutiny, and decreases in skills. Japan and other countries found similarities in these points as well.

    Second, since AI technology has not had a noticeable impact on employment, wages, and working conditions in many cases, collective consultations between labor and management on AI technology have not been conducted. This is another similarity between the situations in Japan and other countries.

    Third, on the other hand, collective labor-management discussions on AI technology were mainly conducted in Germany and Austria. This was not the case in Japan, showing differences between Japan and other countries.

    Fourth, informal communication between labor and management was carried out in the form of worker involvement in the development process of AI technology and briefings for workers on the development and introduction of AI technology. In this point, similarities were found between the Japanese cases and those of other countries.

    Fifth, informal labor-management communication in Japan was seen as "more systematic" than in other countries, suggesting that it could be a Japanese characteristic.

    It was also suggested that the nature of labor-management relations was related to the employment, wages, reallocation, education and training, and the function of AI technology, respectively.

Reference

Milanez, Anna. 2023. "The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation." OECD Social, Employment and Migration Working Papers no. 289, OECD, Paris.

Policy Implications

The results of the survey suggested below respectively:

  1. AI technology can be one of the options to cope with labor shortages due to population decline;
  2. When conducting skill development related to AI technology, it is necessary to create a plan that targets specific groups of people and addresses the needs of older workers;
  3. When considering labor mobility among firms, personnel with specialized AI skills should have multiple skills and knowledge, i.e., specialized skills and knowledge of business areas; and
  4. In collective bargaining, sufficient information sharing and knowledge of AI technology on the part of workers are important.

Research Period

FY2023

Author

IWATSUKI Shinya
Researcher, The Japan Institute for Labour Policy and Training

Contents

JILPT Research Report No.228, full text (PDF:1.6MB) [in Japanese]

Category

Employment/Unemployment

For Citation

JILPT. 2023. Shokuba ni okeru eiai gijutsu no katsuyo to jugyoin e no eikyo; OECD to no kokusai hikaku kenkyu ni motozuku nihon no ichizuke [AI technology implemented in the workplaces and its impact on workers: Japan's position based on an international comparative study with the OECD]. JILPT Research Report no.228. Tokyo: The Japan Institute for Labour Policy and Training.

Related Research

  • JILPT. The Impact of the AI Technology on the Workplace in the Financial Sector: Joint research with OECD. JILPT Research Material Series, No.253. (2022)
  • The Impact of AI Technology on the Workplace in the Manufacturing Sector: Joint Research with OECD. JILPT Research Material Series, No.262. (2023)

JILPT Research Report at a Glance

GET Adobe Acrobat Readeropen a new window To view PDF files, you will need Adobe Acrobat Reader Software installed on your computer.The Adobe Acrobat Reader can be downloaded from this banner.
"e-Letter Japan Labor Issues" Sign Up Now!

AltStyle によって変換されたページ (->オリジナル) /