This chapter examines how AI affects job quantity and skill requirements in Japan. As in other countries, there is currently no clear evidence that AI technologies are significantly reducing job quantity or rendering workers’ skills obsolete. In fact, AI-induced job loss may be less common in Japan than in other countries due to its unique long-term employment practices and chronic labour shortages driven by demographic change. Focusing on the next ten years, many Japanese workers – particularly those in high-skilled occupations – believe AI is more likely to generate new employment opportunities than cause job loss. They also anticipate significant changes in the skills required in their current jobs. In contrast, among AI users in low-income groups and non-regular jobs, concern about job loss outweighs expectations for job creation. There are also regional disparities in the expected impact of AI on job quantity. This chapter’s findings highlight the importance for Japanese policymakers to implement effective active labour market policies (ALMPs) to support workers through these changes.
Artificial Intelligence and the Labour Market in Japan
Abstract
This chapter analyses the impact of AI on job quantity and skills needs by examining both the current situation and future expectations regarding these impacts among Japanese workers.
In Japan, workers in companies that have adopted AI are less likely than their counterparts in other countries to know colleagues who have lost jobs due to AI. This may reflect Japan’s unique employment system – hiring into companies rather than into specific roles, with regular HR-led internal rotations under long-term employment practices. Labour shortages, intensified by a declining birth rate and an ageing population, may also explain why AI-induced job losses appear less common in Japan than in other countries.
Workers in Japanese companies that have adopted AI appear more concerned about job loss over the next ten years than their counterparts in other countries. This may reflect the situation highlighted in Chapter 1: while Japan’s current AI adoption rate remains low by international standards, a high proportion of workers in AI-adopting companies expect its use to increase over the next decade, potentially resulting in more job loss. In addition, it is possible that Japanese workers, because of the lower adoption rate in Japan, are less familiar with the technology and its impact than workers in other countries.
AI will not only destroy jobs but also create them. Among Japanese AI users, the proportion expecting job creation from AI over the next ten years exceeds the proportion expecting job loss. Japanese AI users may believe that AI-driven improvements in their performance will support company growth, and that better working conditions will help them stay in their organisation for longer. In contrast, among workers in companies that have not adopted AI, the proportion expecting job creation is lower than the proportion expressing concern about job loss.
In Japan, AI users who are female, younger, in regular employment, without a university degree, with a disability, balancing work with caregiving responsibilities, and are employed in companies with less than 1 000 employees, are more likely to expect both job creation and job loss from AI, while still maintaining an optimistic outlook in which job creation outweighs job loss. However, among Japanese AI users in low-income groups and in non-regular employment, worries about job loss outweigh expectations of job creation, despite the fact that they report a positive impact of AI on job performance and working conditions. One possible explanation is that these workers tend to spend a large portion of their working time on routine and repetitive tasks and are generally more vulnerable to employment adjustments, which make them more likely to envisage a future in which their work is replaced by AI.
The share of Japanese AI users expecting job creation/loss varies considerably across occupations. Those working as Plant and machine operators, and assemblers or in Elementary occupations, are more likely to expect significant impacts on job quantity due to AI, with job loss expected to outweigh job creation. On the other hand, Managers and Professionals are more likely to expect job creation than job loss. In general, even in occupations where job creation is more prevalent, transitioning from low-skilled to high-skilled work – such as managerial or professional positions – is not straightforward.
The share of AI users expecting job creation/loss also varies considerably across regions. AI users living in Kyushu/Okinawa expect the greatest impact on job quantity over the next ten years and are the most optimistic about job creation. In contrast, AI users in Kinki expect job loss to outweigh job creation during the same period. In general, some workers may face difficulties relocating to regions with higher levels of job creation.
Japanese AI users, like their counterparts in other countries, tend to perceive AI as complementing their skills rather than making them obsolete. However, many Japanese workers expect AI will change the abilities and skills required over the next ten years. The proportion of AI users anticipating significant changes in skill requirements in their current occupation tends to be higher in occupations such as Managers, Professionals, and Technicians and associate professionals.
Japanese AI adopters anticipate that, over the next decade in the AI era, some skills will become increasingly important, such as: the ability to identify and solve problems, adaptability to change, and ethics and compliance awareness.
Japanese AI users who have access to company-based training and financial support to help workers to work with AI, those who are consulted about the introduction of new technologies in the workplace, those whose companies have established internal rules or guidelines for working with GEAI, and those who trust that their employers use only safe and trustworthy AI, are more likely to believe that AI will induce changes in job quantity (with expectations around job creation exceeding those of job loss) and skill requirements of current occupations over the next ten years.
To maximise the benefits of using AI in the workplace, Japanese policymakers may wish to consider:
Enhancing the matching functions of local PES (known as Hello Work in Japan) and actively providing vocational training for individuals seeking re‑employment. The use of AI at Hello Work should also be promoted to improve job-matching and the working conditions of staff. Some workers may face difficulties in moving to high-skilled occupations or relocating to other regions. Therefore, in the future, it may be useful to consider support measures that create new employment opportunities in fields where jobseekers can make the most of their existing human capital, in order to prevent declines in their wages or working conditions.
While Japanese AI users in high-skilled occupations are more likely to expect changes in skill requirements due to AI, such changes are also expected in low-skilled occupations. It is therefore important to strengthen support for reskilling and upskilling in ways that are tailored to the distinct needs of each of these groups. For low-skilled workers, improving equitable access to training opportunities is essential, including through financial support for training costs. For high-skilled occupations, which are generally characterised by heavy workloads, it may be more effective to enhance the accessibility of training in ways that accommodate their working patterns – such as expanding opportunities for online training – and to improve course content to better reflect their needs, rather than focussing primarily on financial support. It is also important to improve awareness among workers of available training opportunities by providing them with better information.
3.1.1. The impact of AI on Job quantity
According to existing studies, AI has so far had no significant negative impact on overall employment. These findings suggest that a jobless future is unlikely; however, workers will need to adapt to the changes brought about by AI.
High exposure to AI does not necessarily imply workers in these occupations will be displaced. Theoretically, there are various channels through which the introduction of AI in the workplace could impact labour demand. Firstly, AI can substitute workers by automating tasks previously performed by human labour (displacement effect). Secondly, as some tasks are automated and AI can complement workers helping them perform tasks more efficiently, productivity increases and costs are reduced. This leads to lower quality-adjusted prices, potentially increasing product demand and, consequently, the demand for workers essential in the production process (productivity effect). Lastly, AI can create new tasks and jobs, particularly in AI development and maintenance (reinstatement effect). Therefore, the overall effect of AI on labour demand is theoretically ambiguous and depends on which effects dominate (Acemoglu and Restrepo, 2019[1]). To understand the impact of AI on aggregate employment empirical research is needed.
So far, there is little evidence of negative aggregate employment outcomes due to AI. Instead, there appears to be a slight positive relationship between AI exposure and employment growth, suggesting that AI may be creating more jobs than it is destroying. At the same time, specific AI technologies could have different, and in some cases negative, impacts. What most studies highlight, is that while more jobs may be impacted by AI, very few are at risk of disappearing entirely. Most occupations involve a combination of skills and abilities that can and cannot be automated. Even highly impacted occupations are unlikely to be fully replaced by automation. Instead, work may need to be organised differently, and workers in these roles may require retraining as technology takes over certain tasks (Lassébie and Quintini, 2022[2]).
Case studies carried out by the OECD in the finance and manufacturing sectors of 8 OECD countries1 in 2022 showed that for 23% of the firms interviewed, AI technologies reduced the number of jobs in the most affected occupations. However, most firms managed these reductions by reallocating workers within the company or through attrition, keeping employees until they either left voluntarily or retired. In addition, firms often opted to slow hiring instead of implementing job cuts, using this approach as a safeguard against the potential failure or underperformance of AI solutions (Milanez, 2023[3]). This is consistent with the finding by Acemoglu et al. (2022[4]) that firms more exposed to AI reduce their overall hiring. A few studies, exploiting variation in AI adoption across U.S. commuting zones, find a negative effect of AI exposure on employment (Huang, 2024[5]; Bonfiglioli et al., 2025[6]).
The majority (77%) of the companies participating in the aforementioned case studies reported no impact on the quantity of jobs for workers most affected by AI technologies. Half of these firms implemented AI technologies to boost production volumes or improve product or service quality, rather than to reduce labour costs. For the other half, the implementation of AI led to the reorganisation of jobs, with workers displaced from certain tasks reassigned to other existing or new tasks. In some cases, AI technologies automated tasks that constituted only a minor share of workers’ jobs, thus not leading to displacement. In other cases, job reorganisation affected more substantial shares of workers’ tasks. However, these jobs were not eliminated, and the automation of certain tasks allowed workers to focus on more complex tasks that could not yet be automated (Milanez, 2023[3]). Additionally, early findings from the OECD SME Survey on Generative AI reveal that, across all countries surveyed, 83% of SMEs report that the use of generative AI has had no effect on the overall number of staff they need (OECD, 2025[7]). Similarly, several studies do not find a significant relationship between AI exposure and aggregate employment (Felten, Raj and Seamans, 2019[8]; Acemoglu et al., 2022[4]; Georgieff and Hyee, 2021[9]). However, it is possible that significant impacts on aggregate economic data only become detectable once the technology is widely adopted and the necessary complementary processes and assets are developed (Brynjolfsson, Rock and Syverson, 2017[10]; Acemoglu et al., 2022[4]; Lane, 2024[11]).
More recent studies by Acemoglu et al. (2023[12]) and Lane (2024[11]) find a small, positive and statistically significant effect of AI exposure on aggregate employment, although direct causality is difficult to prove. The positive association between AI exposure and employment could be due to a productivity effect, or because AI creates new jobs directly. Green and Lamby (2023[13]) find that employment growth for the AI workforce, defined as workers with the skills necessary to develop and maintain AI systems, is strong. On average employment growth was 63% for the AI workforce between 2017 and 2019 – although this workforce is still relatively small, representing less than 0.3% of workers overall. Acemoglu et al. (2022[4]) also find a rapid take‑off of AI vacancy postings starting in 2010 and significantly accelerating around 2015‑2016. Moreover, in 30% of the OECD case studies, interviewees noted that employment was increasing in occupations related to the development and maintenance of AI (Milanez, 2023[3]). In Japan as well, a simulation using a DSGE‑type macroeconomic model to analyse the impact of generative AI on the Japanese economy suggests that the reduction in unemployment outweighs the increase, resulting in a net positive effect on aggregate employment (Tanaka and Nitta, 2024[14]). Their simulation also indicates that employment among workers in occupations susceptible to substitution by generative AI is likely to be negatively affected, suggesting a potential increase in labour market polarisation. However, it also highlights that this negative impact could be reversed through effective policy measures such as reskilling initiatives.
The impact on employment may vary on the type of technology. Technologies like Industrial Automation, Intelligent Logistics, Digital Advertising, and Machine Learning appear to have a negative impact on overall employment, while Medical Information and Autonomous Vehicles have no effect, and E‑Learning, Workflow Management, and Medical Imaging have a positive one (Prytkova et al., 2024[15]).
3.1.2. The impact of AI on skills
The demand for AI skills in the labour market is increasing. Using data on skill requirements in online vacancies, Alekseeva et al. (2021[16]) show that the demand for these skills quadrupled over the period 2010 to 2019. The skills most demanded in AI vacancies are machine learning, natural language processing, deep learning, image processing, programming languages like Python, and big data management (Alekseeva et al., 2021[16]; Manca, 2023[17]; Squicciarini and Nachtigall, 2021[18]). These skills will be needed not only to design the algorithms, but also to explain their functioning to nontechnical professionals, and to monitor outcomes to make sure that AI systems are operating as intended, detecting mistakes and potential biases, and addressing any unintended consequences (Wilson, Daugherty and Morini-Bianzino, 2017[19]). Job postings that require specialised AI skills also necessitate high-level cognitive skills such as creative problem solving, social skills and management skills (project and people management), suggesting that these skills are complementary. Conversely, these jobs typically do not require routine skills, like general administrative and clerical tasks. As a result, an increase in AI-related employment is likely to drive demand for high-level cognitive skills while decreasing demand for routine skills (Alekseeva et al., 2021[16]; Manca, 2023[17]).
Nonetheless, most workers who will interact with pre‑developed AI products may not need AI-specific skills or a deep understanding of AI systems. A survey of AI start-ups found that only 10% required users of their AI products to have expert coding or data skills, while 59% required only general computer familiarity, and the remainder required no special skills at all (Bessen et al., 2023[20]).
In many cases, AI adoption has not yet significantly changed skill requirements within firms. In 2022, 57% and 48% of firms that had adopted AI in finance and manufacturing, respectively, reported no change in skill needs (Lane, Williams and Broecke, 2023[21]). Similarly, in case studies of firms having implemented AI in finance and manufacturing, 60% of firms said that AI adoption had not yet modified skill requirements (Milanez, 2023[3]). This could be partly because AI adoption at the time of those studies was still relatively low and many firms were only experimenting with the technology, but also because interacting with AI applications often requires only basic digital skills, such as the ability to use a computer or smartphone, relying on existing skills.
That being said, 40% of the firms interviewed as part of the above‑mentioned case studies reported a need for new skills, including specialised AI skills and analytical skills. As simple tasks become automated, the proportion of complex tasks performed by workers rises, necessitating specialised knowledge and advanced analytical skills, such as the ability to comprehend and apply new ideas (Milanez, 2023[3]). Managers using algorithmic management software report that the use of such tools increases the most managers’ need for the ability to use or interpret data, and for digital skills (Milanez, Lemmens and Ruggiu, 2025[22]). Employers also say that, while AI has increased the importance of specialised AI skills, it has increased the importance of human skills, such as creativity and communication, even more, as well as the need for highly educated workers (Lane, Williams and Broecke, 2023[21]). Green (2024[23]) shows that occupations with high AI exposure predominantly demand management, business processes, social and digital skills, with the largest increase in demand for skills related to collaboration, originality, and basic office tools. In Japan, an expert panel had highlighted a range of essential skills required for human resources to promote digital transformation in the era of generative AI. These include a willingness to embrace change and a commitment to continuous learning; digital literacy, encompassing ethical awareness and a systematic understanding of knowledge; and, in the context of prompt engineering, the abilities to formulate meaningful questions, develop and test hypotheses, and critically evaluate and select appropriate outputs (METI, 2024[24]).
This section examines workers’ worries about job loss and expectations regarding job creation over the next ten years. The section then analyses the impact of AI on tasks and skill needs.
3.2.1. The expected impact of AI on job quantity
Japanese AI adopters appear less likely than those in other countries to know of individuals in their company who have lost jobs due to AI (Figure 3.1). In the finance and insurance sector, the proportion of Japanese AI adopters who report knowing individuals in their company who had lost their jobs due to AI is 12.9%, which is lower than the 20.4% observed in the other surveyed countries. In the manufacturing sector, the corresponding figure for Japan is 15.0%, which is broadly in line with the 14.7% observed in the other countries. This apparent smaller impact of AI on job loss may reflect Japan’s unique employment system where individuals are hired into companies rather than specific jobs and are regularly reassigned through HR-led internal rotations under long-term employment practices.
Furthermore, AI-induced job loss may also be linked to the extent of labour shortages faced by companies. Among Japanese AI adopters who assess their company as experiencing labour shortages, 16.7% report knowing individuals in their company who had lost their jobs due to AI. By comparison, the corresponding figure among AI adopters who perceive their company as having an appropriate labour supply or labour excess is 20.3%. Therefore, given Japan’s distinctive employment system – which remains prevalent across many Japanese companies – and, in addition, its demographic context characterised by a declining birth rate and an ageing population, which increases the likelihood of labour shortages, it is possible that job losses resulting from AI adoption in the workplace may be less common in Japan than in other countries.
That said, the JILPT survey also asked AI non-adopters whether they know of individuals at other companies within the same sector who have lost their jobs due to AI. At the aggregate level across all sectors in Japan, 4.3% of workers report knowing individuals either at their own company (AI adopters) or at another company in the same sector (non-adopters) who have lost their job due to AI. Although this is a relatively small share, these findings suggest that some Japanese workers have already been affected by job loss due to AI.
Percentage of AI adopters
Note: AI adopters were asked: “Do you know of anyone in your company who has lost their job because of AI?”
Source: OECD worker survey on the impact of AI on the workplace (2022), JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Turning to worries about future job loss, Japanese workers who work in companies that have adopted AI appear more concerned about job loss over the next ten years than those in other countries. In the finance and insurance sector, 73.8% of Japanese AI users are worried, compared to 68.0% in other countries, while in the manufacturing sector the equivalent shares are 71.3% in Japan versus 65.8% in other countries (Figure 3.2). This may reflect the situation highlighted in Chapter 1: while Japan’s current workplace AI adoption rate remains low by international standards, a high proportion of workers in AI-adopting companies expect its use to increase over the next decade, which may contribute to comparatively greater concerns about job loss, compared to other countries.
Percentage of all employees, by whether they and their company use AI
Note: All employees were asked: “How worried are you about losing your job as a result of AI in the next 10 years?”
Source: OECD worker survey on the impact of AI on the workplace (2022), JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
However, it is important to consider the impact of AI on job quantity not only in terms of job losses but also in terms of job creation. Indeed, among Japanese AI users, the proportion expecting job creation from AI exceeds the proportion expressing worries about job loss, both over the next two years and the next ten years (Figure 3.3). Among AI non-users, the proportion expecting job creation from AI exceeds the proportion expressing worries about job loss over the next two years, but not over the next ten years. In contrast, among AI non-adopters, the proportion expecting job creation from AI is lower than the proportion expressing worries about job loss over both the next two and ten years. One likely reason why Japanese AI users are more optimistic about AI’s impact on job quantity is the broadly positive evaluation of AI’s effects on job performance and working conditions across a wide range of workers, as identified in Chapter 2. Japanese AI users may expect that the improvements in their own job performance through the use of AI may contribute to company growth, and that enhancements in various aspects of the working environment will make it easier to continue working within their organisation – leading them to believe that job creation from AI is more likely than job loss.
Percentage of all employees, by whether they and their company use AI
Note: All employees were asked: “How worried are you about losing your job as a result of AI in the next 2 years / in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 2 years / in the next 10 years?” The figure of “Worries about job loss” shows the proportion of employees who answered that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of employees who answered that job creation will be (very or extremely) expected by AI.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Worker characteristics and the expected impact of AI on job quantity
Expectations about job loss and job creation vary across workers. Female AI users, younger ones, those in regular employment, those without a university degree, those with disabilities, and those balancing work with caregiving responsibilities are more likely to expect both job creation and job loss from AI, while maintaining an optimistic outlook in which job creation is expected to outweigh job loss (Figure 3.4). However, among AI users in low-income groups and non-regular employment, worries about job loss outweigh expectations for job creation, despite these workers reporting positive assessments of AI-driven improvements in job performance and working conditions.2 One possible explanation is that, while these workers are currently experiencing positive effects from AI, their engagement in routine and repetitive tasks for much of a typical workday may lead them to believe their tasks could be replaced by AI. Furthermore, compared to regular employees, non-regular workers are generally more vulnerable to employment adjustments. This may reinforce their pessimism regarding the future impact of AI on job quantity. These findings suggest the need to consider, where necessary, support for job creation tailored to the needs of these jobseekers in the future. This should be complemented by well-designed training opportunities to help them acquire the necessary skills for emerging occupations.
Percentage of AI users, by gender, age, employment status, education, annual income in 2023, disability and care status
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” The figure of “Worries about job loss” shows the proportion of AI users who answered that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of AI users who answered that job creation will be (very or extremely) expected by AI. The figure for “University degree” is the sum of four‑year university and graduate school. The annual income figures for 2023 are before taxes and social security contributions have been deducted. “Low” is classified as Below JPY 2 000 000. “Middle” is classified as over JPY 2 000 000 and below JPY 8 000 000. “High” is classified as over JPY 8 000 000.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Company size and the expected impact of AI on job quantity
In general, AI users in smaller companies expect more job change than those in larger companies. However, even in smaller companies, users are more likely to expect AI to result in job creation than job loss (Figure 3.5). The combined proportion of AI users reporting worries about job loss and expectations for job creation is 79% among those working in SMEs, 85% for those in companies with 301999 employees, 64% for those in companies with 1 000‑9 999 employees, and 48% for those in companies with 10 000 or more employees. Furthermore, the D.I. (Job creation − Job loss) score is 4.6 p.p. for AI users in SMEs, 6.9 p.p. for those in companies with 301‑999 employees, 3.5 p.p. for those in companies with 1 000‑9 999 employees, and 4.8 p.p. for those in companies with 10 000 or more employees. This suggests that AI users in smaller companies may face greater changes in job quantity due to AI, but job creation is still expected to outweigh job loss.
Percentage of AI users, by company size
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” The figure of “Worries about job loss” shows the proportion of AI users who answered that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of AI users who answered that job creation will be (very or extremely) expected by AI.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Occupation and the expected impact of AI on job quantity
Expectations regarding job loss and job creation vary considerably across occupations. In particular, AI users working as “Plant and machine operators, and assemblers” or in “Elementary occupations” are more likely to expect significant impacts on job quantity due to AI, with job loss expected to outweigh job creation (Figure 3.6). On the other hand, “Managers” and “Professionals” are more likely to expect job creation than job loss. In general, even in occupations where job creation is more prevalent, transitioning from low-skilled work to high-skilled roles – such as managerial or professional positions – is not necessarily straightforward. These findings suggest that, over the next decade in Japan, in certain occupations where job loss is expected to outpace job creation, it will be important to facilitate the likelihood of re‑employment in occupations with stronger job creation prospects by actively providing vocational training through Public Employment Services (PES). In addition, where necessary, considering targeted job creation support in fields where jobseekers can make use of their existing human capital could be a viable option to improve re‑employment opportunities in the future.
Looking at occupations in more detail, the top 10 occupations where the share of employees expecting job creation exceeding job loss the most, are:
Numerical clerks (40.0 p.p.)
Nursing and midwifery professionals (23.1 p.p.)
Legal professionals (20.5 p.p.)
Engineering professionals (excluding electrotechnology) (20.3 p.p.)
Business services and administration managers (15.6 p.p.)
Sales and purchasing agents and brokers (14.3 p.p.)
Electrotechnology engineers (12.5 p.p.)
Software and applications developers and analysts (12.2 p.p.)
Sales, marketing and development managers (10.0 p.p.)
Mining, manufacturing and construction supervisors (9.7 p.p.)
By contrast, the 5 occupations where the share of employees expecting job loss exceeds the share expecting of job creation the most, are:
Primary school and early childhood teachers (13.0 p.p.)
Protective services workers (4.8 p.p.)
Other health professionals (4.6 p.p.)
Other sales workers (4.0 p.p.)
General office clerks (3.9 p.p.)
Percentage of AI users, by occupation
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” The figure of “Worries about job loss” shows the proportion of AI users who said that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of AI users who said that job creation will be (very or extremely) expected by AI. Occupations of AI users with fewer than 20 samples were omitted.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
This box examines how 22 000 employees in Japan who responded to the JILPT survey assess the impact of AI on job quantity in their own occupations over the next ten years. Based on ISCO digit 3, this analysis summarises, for occupations with 50 or more respondents, the share of employees who said they were “very” or “extremely” worried about job loss due to AI, as well as the share of those who expected that AI was “very” or “extremely” likely to result in job creation (Annex Table 3.A.1).
The top 10 occupations with the highest combined share of employees reporting worries about job loss and expectations of job creation due to AI – indicating a greater likelihood of significant job change – are:
Administration professionals (76.7%, predominance of job creation)
Administrative and specialised secretaries (71.4%, predominance of job loss)
Finance professionals (49.4%, predominance of job creation)
Street and market salespersons (45.0%, predominance of job loss)
Database and network professionals (44.3%, predominance of job loss)
Sales and purchasing agents and brokers (44.3%, predominance of job creation)
Information and communications technology service managers (40.0%, job creation equals job loss)
University and higher education teachers (39.5%, predominance of job loss)
Legal professionals (35.7%, predominance of job creation)
Social and religious professionals (34.0%, predominance of job loss)
The top 10 occupations with the largest positive gap, where the share of employees reporting expectations of job creation due to AI exceeds the share reporting worries about job loss, are:
Medical doctors (14.5 p.p.)
Market gardeners and crop growers (9.5 p.p.)
Building frame and related trades workers (5.6 p.p.)
Retail and wholesale trade managers (4.5 p.p.)
Paramedical practitioners (3.8 p.p.)
Engineering professionals (excluding electrotechnology) (3.0 p.p.)
Armed forces occupations (3.0 p.p.)
Architects, planners, surveyors and designers (3.0 p.p.)
Manufacturing, mining, construction, and distribution managers (2.8 p.p.)
Secondary education teachers (2.8 p.p.)
The top 10 occupations with the largest positive gap, where the share of employees reporting worries about job loss due to AI exceeds the share reporting expectations of job creation, are:
Locomotive engine drivers and related workers (25.7 p.p.)
Street and related service workers (14.7 p.p.)
Keyboard operators (13.2 p.p.)
Professional services managers (12.1 p.p.)
Administrative and specialised secretaries (10.7 p.p.)
Client information workers (10.7 p.p.)
Creative and performing artists (10.3 p.p.)
Sheet and structural metal workers, moulders and welders, and related workers (10.2 p.p.)
Hairdressers, beauticians and related workers (9.8 p.p.)
Manufacturing labourers (9.0 p.p.)
Overall, while it is not possible to determine precisely which types of AI technologies each respondent had in mind when responding to the survey, the fact that job creation is expected to outweigh job loss in certain occupations may reflect several possibilities. These include: (1) that in fields such as healthcare, where AI can support human judgement and help prevent human error, an increased sense of reassurance among prospective workers may lead to a rise in labour supply; (2) that AI could improve efficiency in planning and design processes – such as those involved in architecture – thereby increasing related labour demand; and (3) that labour demand may rise for occupations that involve managing the implementation and use of AI, or that require specialised expertise in AI. Based on such evaluations, respondents may have judged that AI is more likely to lead to job creation in these occupations.
The expected impact of AI on job quantity across regions
The balance between the proportion of AI users reporting worries about job loss and those expecting job creation varies across regions (Figure 3.7). The combined proportion of AI users reporting job loss and job creation is highest in Hokkaido/Tohoku (89%), followed by Kyushu/Okinawa (73%), South Kanto (69%), Hokuriku/Tokai (64%), Kinki (62%), North Kanto/Koshin (59%), and Chugoku/Shikoku (50%). Furthermore, the D.I. (job creation − job loss) score is highest among AI users in Kyushu/Okinawa (13.9 p.p.), followed by Chugoku/Shikoku (5.4 p.p.), South Kanto (5.1 p.p.), Hokkaido/Tohoku (4.8 p.p.), Hokuriku/Tokai (2.0 p.p.), North Kanto/Koshin (0.0 p.p.), and Kinki (‑1.4 p.p.). In general, some workers may face difficulties relocating to regions with higher levels of job creation. Therefore, in regions where job loss is expected to outpace job creation, it is important for the national government and local authorities to collaborate in promoting local job creation aligned with employer demand. At the same time, strengthening the matching functions of the PES will be essential in fostering an environment that facilitates re‑employment.
Several hypotheses were examined to explain why AI users residing in the Kinki region are more likely to expect job loss than job creation.3 The findings suggest that this result may lie in differences in perception stemming from the nature of the AI technologies currently in use in the Kinki region.
Percentage of AI users, by residential area
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” The figure of “Worries about job loss” shows the proportion of AI users who said that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of AI users who said that job creation will be (very or extremely) expected by AI.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
3.2.2. The impact of AI on the volume of work tasks
This section will examine changes in task composition before and after the implementation of AI. Specifically, responses from AI users to the questions: “Thinking about your job, has AI automated any tasks that you used to do?” and “Thinking about your job, has AI created new tasks that you did not do previously?” will be analysed.
In Japan as well as in other countries, the proportion of AI users reporting task automation tends to be higher than those reporting task creation, which suggests that the current use of AI in the workplace is focussed more on simplifying or streamlining existing tasks than on assigning workers new types of responsibilities. At the same time, fewer AI users in Japan report task creation or task automation than in other countries (Figure 3.8). In the financial and insurance sector, 45.4% of Japanese AI users report task creation due to AI, whereas 58.6% do so in other countries. Similarly, 58.5% of AI users in Japan’s financial and insurance sector report task automation, compared to 80.6% in other countries. In the manufacturing sector, the figures are 38.3% for Japan and 46.9% for other countries. Likewise, 57.2% of AI users in Japan’s manufacturing sector report task automation, compared to 70.2% in other surveyed countries. Furthermore, the proportion of AI users reporting both task creation and automation is lower in Japan. In the financial and insurance sector, 36.2% of Japanese AI users report both changes, compared to 50.3% in other countries. In the manufacturing sector, the figures are 26.8% for Japan and 35.5% for other countries. Additionally, the proportion of AI users reporting no occurrence of either task creation or task automation is higher in both sectors in Japan than in other countries. In the financial and insurance sector, 20.0% of Japanese AI users report neither change, compared to 8.9% in other countries. In the manufacturing sector, the figures are 26.6% for Japan and 15.8% for other countries.
Percentage of AI users
Note: AI users were asked: “Thinking about your job, has AI automated any tasks that you used to do?”, “Thinking about your job, has AI created new tasks that you did not do previously?”
Source: OECD worker survey on the impact of AI on the workplace (2022), JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
As further context, across all sectors in Japan, 42.1% of AI users report experiencing task creation due to AI implementation, while 58.6% report task automation. Additionally, 31.1% report experiencing both changes, and 22.8% neither.
Two possible explanations for the fact that many Japanese AI users report no change in their tasks despite the introduction of AI technologies are as follows. First, while Japan’s overall rate of AI adoption remains low, the proportion of AI technologies in use that consist of AI-equipped robots may be comparatively higher than in other countries due to severe labour shortage. These robots are typically deployed to carry out tasks that had previously remained unperformed due to staffing shortages and are generally expected to have little impact on the tasks of existing employees. For example, in some Japanese hotels, robots equipped with AI technologies are deployed at reception desks due to difficulties in securing front desk staff.4 In the security industry, AI-driven security robots are being utilised in response to shortages of security personnel.5 In the food delivery sector, AI-enabled delivery robots are increasingly being used to compensate for the lack of available couriers.6 Similarly, in the food service sector, AI-equipped robots are actively providing meal delivery services to customers as a substitute for human staff. Given Japan’s rapidly ageing population and the already severe labour shortages, it is possible that AI adoption in the workplace tends to focus more on alleviating labour gaps – particularly through the deployment of AI-enabled robots to take over specific tasks. A second possible explanation is that, even when AI technologies have been introduced into the workplace, some companies may have failed to realise meaningful effects from their adoption.
Finally, when looking at the types of tasks automated/created by AI, Japan may be lagging behind other surveyed countries in automating repetitive and complex tasks (Figure 3.9). These findings suggest that, compared to other countries, there may still be considerable room for Japan to advance task transformation through AI. As shown in Figure 2.3, task transformation through AI is likely to be a crucial factor in improving job quality. While Japanese companies are generally expected to aim for task transformation through AI adoption, it is important that they actively focus on making this a reality.
Percentage of AI users, by characteristics of tasks
Note: AI users who answered that AI had created or automated tasks in their jobs were asked: “Were most of these tasks repetitive?/complex?/dangerous?”
Source: OECD worker survey on the impact of AI on the workplace (2022), JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
3.2.3. Prospective impact of AI on the skill needs
This section examines the impact of AI on skill requirements. Specifically, it explores: (1) how AI technologies are affecting the skills used in AI users’ current jobs; (2) how AI is expected to influence the skills and abilities required in their jobs in the future; and (3) which types of skills and abilities are anticipated to become more important, assuming that the adoption of AI in the workplace continues to increase.
In both the financial and insurance sector and the manufacturing sector, AI users in Japan – like those in other surveyed countries – tend to agree more with the statement that AI complements their skills than with the view that it makes them obsolete (Figure 3.10). In particular, the proportion of Japanese AI users who agree that AI has made some of their skills obsolete is lower than that observed among AI users in other countries. This may be partly explained by the finding in Figure 3.9, which suggests that task transformation through AI is progressing at a slower pace in Japan than in other countries. However, the fact that a certain proportion of Japanese AI users already agree that their skills have become obsolete highlights the need to strengthen support for reskilling and upskilling in the future.
Percentage of AI users
Note: AI users were asked: “Please think about the skills you need in your job. Do you agree or disagree with the following statements? AI has made some of my skills less valuable/ AI complements my skills/ I have specialised AI skills, such as those needed to maintain or develop AI/ I am enthusiastic to learn more about AI”
Source: OECD worker survey on the impact of AI on the workplace (2022), JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
The proportion of AI users in Japan who agree that they possess specialised AI skills – such as those required to maintain or develop AI – is lower than that in other countries. This may indicate that Japanese workplaces are less likely to offer roles or opportunities requiring such skills, or that workers have had fewer opportunities to acquire them, owing to limited access to relevant training. In order for Japan to achieve a balanced approach between promoting the development and use of AI and strengthening its response to AI-related risks, it will be particularly important – perhaps more so than in other countries – to strengthen efforts to cultivate a workforce with specialised knowledge of AI. While the proportion of Japanese AI users who express enthusiasm to learn more about AI is slightly lower than in other countries, 59% still report such motivation. This indicates a strong latent interest among workers in acquiring AI-related knowledge, underscoring the need to enhance the availability of accessible and effective learning opportunities.
The expected impact of AI on skill needs
There is a clear expectation that skills needs will further change in the future (Figure 3.11). Among AI users, 47.6% think that changes in skill requirements of their current occupation within the next two years are “very” or “extremely” likely. This share increases to 58.1% over the next ten years. Among AI non-users, these shares are 27.6% and 37.4%, respectively, while they are 10.2% and 14.9% respectively amongst non-adopters. Furthermore, when including those who anticipate “moderate” and “slight” changes, these figures rise to 89.0% and 92.5%, respectively. So even among employees who are not currently using AI, there is growing recognition of AI as a transformative force in the workplace.
Percentage of all employees, by whether they and their company use AI
Note: All employees were asked: “What impacts do you think AI will have on the skills or abilities needed for your current job in next 2 years/in next 10 years?”
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
The share of AI users in Japan anticipating that changes in skills requirements within the next ten years will be either “very” or “extremely significant” tends to be higher among higher-skilled occupations (Figure 3.12). For example, 64.7% of “Managers” and 62.9% of “Professionals” expect substantial change. By contrast, this figure drops below 50% for occupations such as “Clerical support workers” (49.3%) and “Elementary occupations” (48.7%). These findings suggest that it is important to strengthen support for reskilling and upskilling in ways that are appropriately tailored to the distinct needs of both high- and low-skilled workers. For the latter, improving equitable access to reskilling and upskilling opportunities is essential, including through financial support to cover training costs. For high-skilled occupations, which are generally characterised by demanding workloads, it may be more effective to enhance the accessibility of training in ways that accommodate their working patterns – such as by expanding opportunities for online training – and to improve course content to better reflect their needs, rather than focussing primarily on financial support. In addition, it is important to strengthen information dissemination to ensure workers are more likely to become aware of available training opportunities.
Percentage of AI users, by occupation
Note: AI users were asked: “What impacts do you think AI will have on the skills or abilities needed for your current job in next 10 years?” The figure shows the proportion of AI users who answered that their skills or abilities required for their current job were changed (extremely significant or very) by AI in the future. Occupations with fewer than 20 samples were omitted.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Skills and abilities in the AI era
The three skills that AI adopters are most likely to expect to increase than decrease are: the “Ability to identify and resolve issues” (41.7 p.p. in the D.I. (Increased – Decreased)), and “Adaptability to change” (39.6 p.p.), followed by “Ethics and compliance awareness” (38.2 p.p.) (Figure 3.13).
A significant share of AI adopters in Japan expect the “Ability to identify and resolve issues” to become more important. This skill is likely viewed as a core capability for human-centred collaboration with AI, in which humans retain a central role in identifying workplace challenges, determining which tasks to delegate to AI, and interpreting AI-generated outputs to devise effective solutions. Similarly, the expectation that “Adaptability to change” will become more important may reflect an awareness of the need for resilience in response to the rapid pace of technological change. As AI systems continue to evolve, workers will be required to adjust swiftly to new tools and workflows. The perceived growing importance of “Ethics and compliance awareness” further suggests that AI adopters are conscious of the risks and complexities inherent in AI use. While AI offers significant benefits in terms of productivity and improved decision making, it also demands a strong ethical foundation.
Percentage of AI adopters
Note: All AI adopters were asked: “Assuming that AI and GEAI will be increasingly adopted in the workplace over the next ten years, how do you expect the following skills and abilities to be affected in your current job?”
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Chapter 2 showed that the following initiatives contributed to maximising the benefits of AI for job performance and working conditions: (1) providing company training and financial support to help employees work effectively with AI, and encouraging employees to engage in self-learning to work with AI; (2) promoting worker consultation on the use of AI technologies; (3) establishing internal rules or guidelines to ensure the appropriate use of GEAI in the workplace; and (4) actively building employee trust – through various forms of communication and the use of external third-party risk assessment bodies – that only safe and trustworthy AI technologies will be introduced. As this section shows, these initiatives not only enhance the impact of AI on job quality, but also makes workers more positive about job creation, while at the same time expecting significant changes in terms of skills requirements.
3.3.1. Training and AI-Driven impacts on job quantity
Japanese AI users who engage in company-provided training and self-learning appear to be more optimistic about job creation while also expecting greater changes to skills requirements of their current occupation. AI users who have not engaged in either company-provided training or self-learning are more likely to say AI will result in job loss than in job creation. They are also less likely to anticipate significant changes in skill requirements of their current occupation (Figure 3.14). AI users who have received both company- provided training and engaged in self-learning, compared to those who have received neither, report a D.I. (Job creation − Job loss) score that is 19.0 p.p. higher, as well as a 37.1 p.p. higher likelihood of expecting substantial changes in skill requirements. This may be because Japanese users who have engaged in both types of learning activities have acquired greater knowledge about AI technologies. Building on this acquired knowledge, it is noteworthy that Japanese AI users are optimistic about job quantity over the next ten years, anticipating that job creation will outweigh job loss.
3.3.2. Worker consultation and AI-Driven impacts on job quantity
When comparing workers who have been consulted about the use of new technologies to those who have not, Japanese AI users who have been consulted by their employers are more likely to expect job creation than job loss, and they are also more likely to anticipate significant changes in skill requirements in their current occupation (Figure 3.15). AI users who have been consulted by their employer, compared to those who have not, report a D.I. (Job creation − Job loss) score that is 7.8 p.p. higher, as well as a 26.6 p.p. higher likelihood of expecting substantial changes in skill requirements. This may reflect the possibility that Japanese users who have been consulted gained a deeper understanding of employers’ perspectives on AI technologies, which in turn heightened their sensitivity to anticipated changes in job quantity and skill requirements. Communicating employers’ perspectives on the use of AI technologies to workers may help Japanese AI users form a more optimistic outlook regarding job quantity – specifically, the expectation that job creation will outweigh job loss over the next ten years.
3.3.3. Internal rules or guideline and GEAI-Driven impacts on job quantity
Japanese GEAI users who indicate that internal rules or guidelines for the appropriate use of GEAI have been established are more likely to expect job creation than job loss, and they are also more likely to anticipate significant changes in skill requirements in their current occupation. GEAI users who report the presence of internal rules or guidelines, compared to those who do not, report a D.I. (Job creation − Job loss) score that is 7.9 p.p. higher, as well as a 20.5 p.p. higher likelihood of expecting substantial changes in skill requirements. This may reflect the possibility that internal rules or guidelines deepen these workers’ understanding of the technology. By adhering to these guidelines, they may envision themselves benefiting from the positive impacts of GEAI. This, in turn, could explain their optimistic outlook regarding job quantity – specifically, the expectation that job creation will outweigh job loss over the next ten years.
Percentage of AI users, by whether they received company training or engaged in self-learning
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” “Has your company provided or funded training so that you can work with AI?(Yes/No/I don't know)” “In 2023, did you engage in reskilling or upskilling to work with AI?(Yes/No/I don't know)” “What impacts do you think AI will have on the skills or abilities needed for your current job in next 10 years?” The figure of “Worries about job loss” shows the proportion of employees who said that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of employees who said that job creation will be (very or extremely) expected by AI. The figure below shows the proportion of AI users who answered that their skills or abilities required for their current job were changed (extremely significant or very) by AI.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Percentage of AI users, by whether their employers consult them regarding the use of new technologies in the workplace
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” “In your experience, does your employer consult workers or worker representatives regarding the use of new technologies in the workplace?” The figure of “Worries about job loss” shows the proportion of employees who said that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of employees who said that job creation will be (very or extremely) expected by AI. The figure below shows the proportion of AI users who answered that their skills or abilities required for their current job were changed (extremely significant or very) by AI.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
3.3.4. Trust in employers and AI-Driven impacts on job quantity
Japanese AI users who trust their company to use AI in a safe and trustworthy way are more likely to report that AI will result in job creation than job loss. They are also more likely to expect significant changes in skill requirements in their current occupation due to AI. AI users who trust their company in this regard, compared to those who do not, report a D.I. (Job creation − Job loss) score that is 14.3 p.p. higher, as well as a 32.6 p.p. higher likelihood of expecting substantial changes in skill requirements. In contrast, Japanese AI users who do not trust their company are more likely to believe AI will result in job loss than in job creation. They are also less likely to anticipate substantial changes in skill requirements in their current occupation over the same period (Annex Figure 3.A.1). These findings suggest that building employees’ trust in their company’s governance and compliance regarding AI technologies fosters a positive outlook among workers that AI will enhance their company’s value and deliver tangible benefits in the future. As a result, Japanese AI users with such trust hold a more positive view of job creation. By contrast, the lack of such trust may foster a negative outlook among workers that AI will undermine their company’s value and put their own jobs at risk.
Overall, the four approaches that maximise the positive impact of AI appear to be associated with more positive views of job quantity as well as greater expectations that skill requirements will change. These findings highlight the importance for Japanese policymakers, when promoting such initiatives, to simultaneously strengthen support for employee reskilling and upskilling, as well as enhance the matching functions of the Public Employment Services (known as Hello Work in Japan) for individuals seeking re‑employment. The Basic Policy on Economic and Fiscal Management and Reform 2025 outlines plans to promote the use of AI at Japan’s Hello Work. In light of this, Japanese policymakers should advance efforts to enhance the matching functions of Hello Work and improve working practices for its staff.
Percentage of all employees
Note: All employees were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” The figure of “Worries about job loss” shows the proportion of employees who said that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of employees who said that job creation will be (very or extremely) expected by AI. Occupations of AI users with fewer than 50 samples were omitted.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
Percentage of AI users, by whether they trust own company to only use AI that is safe and trustworthy
Note: AI users were asked: “How worried are you about losing your job as a result of AI in the next 10 years?” “To what extent do you expect AI will increase employment in your occupation in the next 10 years?” “To what extent would you trust your company to only use AI that is safe and trustworthy?” The figure of “Worries about job loss” shows the proportion of employees who said that job loss will be (very or extremely) worried by AI. The figure of “Expectations for job creation” shows the proportion of employees who said that job creation will be (very or extremely) expected by AI. The figure below shows the proportion of AI users who answered that their skills or abilities required for their current job were changed (extremely significant or very) by AI.
Source: JILPT worker survey on the impact of the introduction of AI into the workplace on working practices (2024).
References
[4] Acemoglu et al. (2022), “Artificial Intelligence and Jobs: Evidence from Online Vacancies”, Journal of Labor Economics, Vol. 40/S1, https://doi.org/10.1086/718327.
[1] Acemoglu, D. and P. Restrepo (2019), “Automation and New Tasks: How Technology Displaces and Reinstates Labor”, Journal of Economic Perspectives, Vol. 33/2, pp. 3-30, https://doi.org/10.1257/JEP.33.2.3.
[12] Albanesi et al. (2023), “New Technologies and Jobs in Europe”, http://www.nber.org/papers/w31357 (accessed on 21 June 2024).
[16] Alekseeva, L. et al. (2021), “The demand for AI skills in the labor market”, Labour Economics, Vol. 71, p. 102002, https://doi.org/10.1016/J.LABECO.2021.102002.
[20] Bessen, J. et al. (2023), “The Business of AI Startups”, SSRN Electronic Journal, https://doi.org/10.2139/SSRN.3293275.
[6] Bonfiglioli, A. et al. (2025), “Artificial intelligence and jobs: evidence from US commuting zones”, Economic Policy, Vol. 40/121, pp. 145-194, https://doi.org/10.1093/EPOLIC/EIAE059.
[10] Brynjolfsson, E., D. Rock and C. Syverson (2017), “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics”, National Bureau of Economic Research Working Paper Series, https://doi.org/10.3386/w24001 (accessed on 25 July 2024).
[8] Felten, E., M. Raj and R. Seamans (2019), “The Occupational Impact of Artificial Intelligence: Labor, Skills, and Polarization”, SSRN Electronic Journal, https://doi.org/10.2139/SSRN.3368605.
[9] Georgieff, A. and R. Hyee (2021), “Artificial intelligence and employment: New cross-country evidence”, OECD Social, Employment and Migration Working Papers, https://doi.org/10.1787/c2c1d276-en.
[23] Green, A. (2024), “Artificial intelligence and the changing demand for skills in the labour market”, OECD Artificial Intelligence Papers, No. 14, OECD Publishing, Paris, https://doi.org/10.1787/88684e36-en.
[13] Green, A. and L. Lamby (2023), “The supply, demand and characteristics of the AI workforce across OECD countries”, OECD Social, Employment and Migration Working Papers, No. 287, OECD Publishing, Paris, https://doi.org/10.1787/bb17314a-en.
[5] Huang, Y. (2024), “The Labor Market Impact of Artificial Intelligence: Evidence from US Regions”, IMF Working Papers, Vol. 2024/199, p. 1, https://doi.org/10.5089/9798400288555.001.
[11] Lane, M. (2024), “Who will be the workers most affected by AI?: A closer look at the impact of AI on women, low-skilled workers and other groups”, OECD Artificial Intelligence Papers, No. 26, OECD Publishing, Paris, https://doi.org/10.1787/14dc6f89-en.
[21] Lane, M., M. Williams and S. Broecke (2023), “The impact of AI on the workplace: Main findings from the OECD AI surveys of employers and workers”, OECD Social, Employment and Migration Working Papers, No. 288, OECD Publishing, Paris, https://doi.org/10.1787/ea0a0fe1-en.
[2] Lassébie, J. and G. Quintini (2022), “What skills and abilities can automation technologies replicate and what does it mean for workers?: New evidence”, OECD Social, Employment and Migration Working Papers, No. 282, OECD Publishing, Paris, https://doi.org/10.1787/646aad77-en.
[17] Manca, F. (2023), “Six questions about the demand for artificial intelligence skills in labour markets”, OECD Social, Employment and Migration Working Papers, No. 286, OECD Publishing, Paris, https://doi.org/10.1787/ac1bebf0-en.
[24] METI (2024), Approaches to the Human Resources and Skills Required for Promoting Digital Transformation in the Era of Generative AI (2024), https://www.meti.go.jp/press/2024/06/20240628006/20240628006-b.pdf (accessed on 3 April 2025).
[3] Milanez, A. (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 Publishing, Paris, https://doi.org/10.1787/2247ce58-en.
[22] Milanez, A., A. Lemmens and C. Ruggiu (2025), “Algorithmic management in the workplace: New evidence from an OECD employer survey”, OECD Artificial Intelligence Papers, No. 31, OECD Publishing, Paris, https://doi.org/10.1787/287c13c4-en.
[7] OECD (2025), “Microdata from the OECD SME Survey on Generative AI”.
[15] Prytkova, E. et al. (2024), “The Employment Impact of Emerging Digital Technologies”, CESifo Working Paper, No. No. 10955, CESifo, Munich, https://www.cesifo.org/en/publications/2024/working-paper/employment-impact-emerging-digital-technologies (accessed on 26 July 2024).
[18] Squicciarini, M. and H. Nachtigall (2021), “Demand for AI skills in jobs: Evidence from online job postings”, OECD Science, Technology and Industry Working Papers, No. 2021/03, OECD Publishing, Paris, https://doi.org/10.1787/3ed32d94-en.
[14] Tanaka, M. and T. Nitta (2024), Econometric analysis of the impact of generative AI on the Japanese economy, https://www.dir.co.jp/report/research/economics/japan/20241108_024729.pdf (accessed on 3 April 2025).
[19] Wilson, H., P. Daugherty and N. Morini-Bianzino (2017), “The Jobs That Artificial Intelligence Will Create”, Vol. 58/4, http://mitsmr.com/2odREFJ (accessed on 21 June 2024).
← 1. The OECD Case Studies of AI Implementation were conducted in Austria, Canada, France, Germany, Ireland, Japan, the United Kingdom, and the United States.
← 2. For AI users in non-regular employment, see Figure 2.7. For low-income AI users, the proportion who reported that each of the following outcomes was improved (either a lot or a little) by AI is as follows: job performance – 58.2%, enjoyment – 38.8%, mental health – 41.4%, physical health – 45.9%, and fairness in management – 45.1%.
← 3. The following hypotheses were not supported by the data: (1) that AI users in these regions are more likely to perceive a labour surplus in their workplace, thereby increasing expectations of job loss; (2) that a higher proportion of AI users in the Kinki region are employed in occupations more prone to anticipating job loss; and (3) that the share of AI users who are non-regular workers or low-income earners – groups typically associated with a stronger tendency to expect job loss – is relatively high.
← 4. For example, “Henn na Hotel,” which offers front desk services provided by robots, operates 21 locations across Japan.
← 5. In Japan, pilot programmes and implementations have begun in airports, commercial facilities, and factories and warehouses.
← 6. METI has been collecting and organising information related to autonomous delivery robots. https://www.meti.go.jp/policy/economy/distribution/deliveryrobot/index.html.