This chapter first reviews the policies and strategies implemented by OECD countries to support upskilling and reskilling for the increasing adoption of AI in the workplace. It highlights examples of publicly-funded AI-related training programmes across OECD countries and explores how countries incentivise employers, adult learning providers, workers, and jobseekers to participate in these initiatives. Finally, it provides an overview of the characteristics of AI-related training provision, based on an analysis of course catalogues from Australia, Germany, Singapore, and the United States.
Training Supply for the Green and AI Transitions
Equipping Workers with the Right Skills
Abstract
Artificial intelligence (AI) is having a profound impact on the world of work. Jobs that require specific AI skills or general AI literacy are on the rise in many countries (Squicciarini and Nachtigall, 2021[1]). At the same time, AI has the potential to make some jobs and tasks redundant, though to date there is limited evidence indicating substantial negative employment effects attributable to AI (OECD, 2023[2]). To support greater adoption of AI-related technologies in the workplace and to enable workers to adapt to these changes, adult learning systems are being called to equip adults with skills both for developing and maintaining AI systems and for adopting and interacting with these systems.
Training for AI varies in terms of its content, the audience it targets and pre‑requisites. Some training courses focus on training high-skilled AI professionals to develop AI models, such as Natural Language Processing or Neural Networks. Other training courses focus on upskilling and reskilling employed or unemployed adults of various backgrounds and skill levels to interact with AI in a work setting, which can include AI literacy. Although lacking an official definition, AI literacy can be described as the comprehension, usage and monitoring of AI applications with critical reflection, without the necessity of being able to develop AI models (Long and Magerko, 2020[3]; Ng et al., 2021[4]). This chapter takes a broad view of training for AI as encompassing the skills and knowledge needed for both general AI literacy courses and courses specifically for AI professionals.
In light of the growing importance and use of AI, many of the 21 countries that responded to the OECD policy questionnaire have introduced AI strategies that outline how they will respond to the opportunities and challenges presented by AI, including considerations for research and innovation, investment, human capital, ethical concerns and international co‑operation. The majority of responding countries have developed national AI strategies dedicated solely to the emergence of AI (Table 4.1). Some countries have instead incorporated AI components into their existing digital strategies. OECD countries that are part of the European Union were encouraged by the European Commission to develop an independent AI strategy or enhance existing strategies with a specific AI component as of 2019. Most strategies place a specific emphasis on education and the development of AI-related and digital skills. Additionally, a couple of countries have also introduced publicly-funded, AI-related training programmes in their policies. This section outlines the different initiatives undertaken by countries.
One category of strategies outlines education measures aimed mainly at increasing the number of AI professionals in the country. AI professionals are highly-skilled individuals who are trained in developing and maintaining AI tools. In most cases this requires investing in formal education by designing, for instance, new master or doctoral studies related to AI. This is the case in France, Japan, Korea, Poland, Sweden and the United States. In France, a two‑phase national strategy for AI with the goal of positioning France as a European and world leader was launched in March 2018. The main focus of the second phase (2022‑25) of France’s AI strategy is the increase of the supply of well-trained AI professionals. A total of EUR 2.22 billion over the next five years will be allocated to this objective, including expanding the national online training offer in data science, AI and robotics. In Korea, the 2019 national AI strategy aims at innovating the education system to increase the number of people working in AI-related fields. Concrete steps to achieve this include the establishment and expansion of AI-related departments and programmes at universities as well as non-degree courses in co‑operation with industry. Action plans which review progress achieved and changing needs will be published on a yearly basis. Singapore’s AI Strategy aims to attract the world’s top AI creators, with the goal of expanding the Singaporean AI Practitioner community to 15 000 individuals. The strategy focuses on enhancing the skills of the workforce through specialised AI training programmes tailored to different sectors.
Another set of AI upskilling and reskilling strategies broaden access to AI training to lower-skilled adults . This is the case in the Czech Republic (Czechia), Germany, Hungary, Lithuania, Singapore, Sweden and in the United Kingdom. In the United Kingdom, alongside rolling out scholarships and fellowships for Masters-level conversion courses in AI and data science, short-duration “Skills Bootcamps” have been established that have low entry requirements. These bootcamps offer free and flexible skills courses aimed at supporting the development of AI-related or digital skills among the broader population. In the Hungarian AI strategy, which was adopted in 2020, one of the key performance indicators is to engage up to 300 individuals holding a PhD in AI-research but also provide AI-related adult education to 8 000 citizens. Strategies in Lithuania and Czechia pay specific attention to workers in occupations with a high likelihood of job task automation, such as the automotive industry.
Some countries have yet to unveil dedicated AI strategies, but have already implemented digital skills strategies, some of which mention AI (see Table 4.1). In the Slovak Republic, AI is mentioned as part of the Strategy of the Digital Transformation of the Slovak Republic 2030, which was published in 2019. Next to reforming the formal education system, the Slovak Republic intends to create AI-related vocational training opportunities targeting both workers and jobseekers. In Austria, in 2019, the Ministry of Finance launched the “Digital Skills Austria” strategy which proposes 350 initiatives and measures to digitalise Austria. One of the eight strategic priorities of the strategy is the diffusion of AI in the world of work.
Each AI policy or strategy provides incentives to key actors in the skills ecosystem to support upskilling and reskilling in AI skills. The next sections highlight some notable government training programmes and describe the types of incentives aimed at workers and jobseekers, employers and adult learning providers.
Note: 1. Initiative is planned to be introduced, but was not yet implemented at the time of the questionnaire response.
Source: 2023 OECD Policy Questionnaire: Adult Learning for Diffusion of AI.
AI-related training programmes
Several OECD countries have introduced publicly-funded, AI-related training programmes to proactively ensure the workforce is well-prepared to leverage the opportunities introduced by this technology (see Table 4.1). These endeavors are still in their infancy and likely need to be scaled up and implemented across OECD countries. AI-related training programmes in this section are classified in two categories: 1) training programmes in general AI literacy, and 2) training programmes to develop AI professionals (see Table 4.2).
Source: 2023 OECD Policy Questionnaire: Adult Learning for Diffusion of AI.
Training programmes in general AI literacy
The pervasive influence of AI across domains underscores the need for the general public to acquire proficiency in utilising AI tools, especially considering the digital problem-solving skills gap. A large share of the adult population is lacking basic digital skills for problem-solving in their day-to-day life (OECD, 2019[5]). Given the complexity of AI technologies, this share could be even higher for problem-solving involving AI. Thus, beyond raising awareness about the emergence, potential and challenges of AI, the general public should also be taught how to efficiently use and interact with AI tools. This could also include training related to the ethics of AI.
Presenting courses in an interactive and dynamic manner, like quests or challenges, is one approach countries are using to engage participants in AI training. In Japan, for instance, the “Manabi Deluxe Quest” is a digital human resources development programme where students and adults learn how to use AI and data science to solve problems in an interactive way. In the first cohort, 96% of the 2 134 participants reported a high satisfaction rate. Research suggests that gamified learning experiences can enhance engagement and retention, ultimately fostering greater proficiency in AI tool utilisation (Hamari, Koivisto and Sarsa, 2014[6]). One example of such a gamified learning experience to raise awareness about the emergence of AI and to teach the general public AI basics is the Hungarian AI challenge set up in 2020. The challenge, which can be completed within two to three hours, was set up by ELTE University and has as its goal to equip 1% of the Hungarian population (100 000 people) with a basic understanding of AI. More advanced courses are supposed to be released in the future. Developed and implemented by the Ministry of Culture and Innovation and higher education institutions, the course was completed by 300 000 Hungarian speakers in the 2022‑23 fall school year, reflecting the challenge’s success in reaching a broad audience.
In Austria, the initiative “Digital Everywhere” (Digital Überall) will roll out 3 500 workshops in all Austrian municipalities throughout 2024 with the aim of bolstering basic digital competencies, including AI and cybersecurity, among the general population. To ensure inclusivity across diverse demographics, the workshops are conducted at diverse venues, including youth centres and retirement homes, facilitating access for individuals from different backgrounds and age groups. Singapore launched a training initiative aimed at assisting jobseekers and employees from various professional backgrounds and skill levels in acquiring digital and AI literacy skills that equip them for the digital economy. The initiative is targeted at adults in jobs likely to be affected by AI and with low levels of skills and education.
Training programmes to develop AI professionals
Given the rising need for highly-skilled AI workers, training AI professionals who can develop and maintain AI tools is an important priority in several OECD countries. Training programmes to develop high-skilled AI professionals sometimes entail the creation of new bachelor, master or doctoral programmes. Poland, as part of the “Academy for Innovative Applications of Digital Technology (AI Tech)” initiative, has co‑ordinated a collaboration between leading companies in the technology sector and Polish universities to jointly create AI-related master programmes. Five universities1 have launched new second cycle programmes and specialisations in AI, cybersecurity and machine learning in 2021, with funding from the European Union.
Some countries have developed short, non-formal training programmes aimed at upskilling or reskilling employed or unemployed adults to cope with changes brought about by the AI transition. In 2020, the United Kingdom Department for Education launched Skills Bootcamps, which offer short sector-specific skills training to both employed and unemployed adults in different areas, including digital and AI. For instance, there is a Skills Bootcamp on Data/AI which helps learners develop the technical and employability skills to secure an entry role as a Data Technician or Analyst. Course curricula is created and delivered in collaboration with employers. The courses last up to 16 weeks and are free of charge for the participant. Upon successful completion of a Skills Bootcamp, participants are offered a job interview with a potential employer, helping workers make the transition from training to work.
Incentives for workers and jobseekers
Incentivising the development of AI skills can be achieved by influencing the demand side, starting with measures targeted to individuals. This section describes whether and how countries are providing financial incentives to steer individuals to undertake training in AI skills, and discusses the role of information, advice and guidance in AI upskilling and reskilling strategies (see Table 4.3).
Source: 2023 OECD Policy Questionnaire: Adult Learning for Diffusion of AI.
Financial incentives for training
Fifteen of the 21 countries that responded to the policy questionnaire report having financial incentives for training which could be used by workers and jobseekers to cover the costs of AI-related skills training. However, none of these financial incentives is explicitly tied to AI skills. All financial incentives are general in nature and could be used towards funding AI-related skills training, though this was not a requirement for their use.
For instance, Switzerland offers induction grants for training programmes via unemployment insurance. These grants cover part of the wage costs during a new employee’s induction phase in order to achieve the level of productivity expected by the employer and to give them time to develop the required skills on-the‑job. Although induction grants are not specifically designed to incentivise workers to develop AI skills, they could be used for that purpose. Czechia offers retraining subsidies up to 82% of the price of the course for digital upskilling and reskilling, including in AI, as part of its “I’m in course” programme. The Public Employment Service of Latvia provides funding of up to EUR 500, with 50% pre‑financing, to unemployed and employed individuals to participate in courses offered by online learning providers. Additionally, unemployed individuals are eligible to receive a stipend of EUR 5 per day. As discussed in Chapter 3, Singapore implemented the SkillsFuture Credit in 2015, providing an opening credit of SGD 500 to Singaporeans aged 25 and above to partially offset training course expenses. Additionally, when Singaporeans turn 40, there is a top-up the credit of SGD 4 000.
In some OECD countries, financial incentives for training are targeted at certain population groups, such as low-skilled individuals, older adults or women. In Australia, for instance, the programme “Skills and Training Incentive” has been set up as an initiative under the More Choices for a Longer Life Package to assist citizens and permanent residents aged 40+ in acquiring new skills and competences to stay in the workforce. After successfully completing the Skills Checkpoint for Older Workers that helps set up a career plan, participants can receive a financial contribution of up to AUD 2 200 for reskilling and upskilling. Training must be related to an occupation in national shortage or with strong or moderate future demand on the National Skills Commission’s Skills Priority List. In order to be eligible, participants must either be employed, at risk of being unemployed or have been unemployed for no more than 12 consecutive months. In 2022‑23, 4 256 people participated in the programme which incurred a total cost of AUD 3.97 million.
Information, advice and guidance initiatives
As the above financial incentives for training do not steer individuals towards developing AI skills, it is up to individuals themselves to make their own training choices. Quality information, advice and guidance is therefore critical for supporting adults in navigating training markets and in understanding which AI skills and knowledge areas are in demand. However, career advice and guidance are not common components of any of the AI policies or strategies reported in the policy questionnaire, as only five countries (Costa Rica, Germany, Norway, Singapore and the Slovak Republic) report having such initiatives in place. Germany has also launched an initiative focused on providing information, advice and guidance, but unlike other programmes aimed at jobseekers and workers, it is designed to benefit employers. Whilst beyond the scope of this study, guidance programmes targeted to employers can also be helpful in the space of AI given the novelty of AI technologies.
Singapore offers career guidance through Skills and Training Advisors who provide tailored advice and training recommendations based on individual needs. The Skills Demand for the Future Economy reports complement this by outlining the key skills required for emerging sectors, such as artificial intelligence. Public employment offices, vocational guidance centres and training providers in the Slovak Republic offer general career counseling and skills development programmes. These services can help individuals navigate the AI transition and adapt to the evolving demands of the workforce. More countries could benefit from expanding their guidance services to cover the delivery of AI-related training and careers information. Germany, although not targeting jobseekers and workers, has created the Civic Coding Innovation Camp which calls on AI-related start-ups/businesses that have ideas on how to use AI for the common good to pitch their projects to AI experts and receive free guidance and counselling. Once selected, the chosen start-ups will receive up to 30 hours of expert counselling and hands-on support from specialists at N3XTCODER and Zukunft Zwei, two renowned German companies specialising in strategy and software development within the tech sector.
Incentives for employers
Firms are at the forefront of the changes induced by the introduction and adoption of AI. Firm-provided training can help employers to harness the opportunities of this new technology, by equipping employees with necessary skills. Larger firms are more likely to have a workforce development strategy and HR and management functions responsible for assessing firm training needs and addressing them (OECD, 2021[7]) (see Box 4.1 for a case study of a large telecommunications firm in Germany providing AI skills training to its employees). Smaller firms face greater challenges than larger firms in covering the cost of training and acquiring the necessary expertise to support the adoption of AI technology. This section identifies four sets of incentives intended to support employers in upskilling and reskilling their employees for AI: 1) subsidies or tax deductions to train employees, 2) employment incentives, 3) apprenticeship programmes, and 4) promotion of private‑public collaborations (see Table 4.4).
Source: 2023 OECD Policy Questionnaire: Adult Learning for Diffusion of AI.
Subsidies or tax deductions to train employees
Ten out of the 21 countries that responded to the policy questionnaire reported having subsidies or tax deductions in place for employers that offer training, though none of these financial incentives have the express purpose of developing AI skills. Going forward, some of these existing incentives can be tailored to encourage participation in AI-specific training.
However, several countries, including Austria, Czechia and Poland, offer subsidies to firms to invest in the development of digital skills. Polish employers who provide training to their employees in priority areas can receive a financial contribution from the National Training Fund. Based on a labour market analysis, Poland defined the training of adults under 30 years of age in the field of digital skills as a priority for 2023. In a similar vein, Austria is offering subsidies to SMEs who provide training aimed at increasing the digital competences of their employees. Funding is provided in the form of non-repayable grants and amounts to a maximum of EUR 5 000 per person. The funding rate can amount to a maximum of 80% of the eligible external training costs for a maximum of 25 people per company. In order to be eligible, the training has to be offered by recognised Austrian educational institutions, such as universities, research facilities or innovation hubs. In Czechia, the government launched two calls in order to financially incentivise employers to provide training to support the development of digital skills. While one call is addressed to the companies themselves, another call is aimed at umbrella employer organisations.
Additionally, there are several countries that offer subsidies to firms to invest in the skills of their employees, but without any effort to steer the content. Countries with such training incentives in place include Germany, Hungary, Japan, Korea, Singapore and the United Kingdom. In Germany, the public employment service offers subsidies to employers who train their employees, with smaller companies receiving larger subsidies relative to their size, and companies with 10 employees or less receiving the full reimbursement. Furthermore, training offered to employees aged 45 and above, low-qualified employees, as well as those with disabilities, is fully subsidised for companies with fewer than 250 employees. Employers can also receive compensation for wages paid to employees while they participate in training. In Hungary, financial assistance is provided for employers who invest in enhancing the skills of their workforce through workplace training. While the programme does not explicitly focus on the advancement of AI skills, it presents an opportunity for workers and companies to better adapt to the evolving demands driven by the proliferation of AI-related technologies through the development of vocational, digital, foreign language and soft skills. Within this initiative, financial support is available for training expenses, wage subsidies, as well as project preparation, management and professional implementation costs. Funding for the programme is provided jointly by the European Social Fund+ and the Hungarian state budget.
Employment incentives
Employment incentives can support individuals at risk of job automation due to the AI transition in remaining in quality positions. These incentives, such as wage subsidies, relocation assistance and support for AI entrepreneurship, can also help firms develop the necessary skills for adopting AI. However, only 5 out of the 21 countries that responded to the policy questionnaire have implemented such employment incentives.
For instance, the Croatian Employment Service (CES) offers enhanced financial support through employment, traineeship and self-employment subsidies for jobs that are primarily related to green or digital activities, including AI. Employers who qualify receive higher subsidies, encouraging the growth of environmentally sustainable and digitally focused activities.
Apprenticeship programmes
A targeted form of employment incentive is a government-funded apprenticeship or traineeship programme. These programmes provide practical, hands-on training that directly addresses the evolving demands of the job market, particularly in sectors aligned with the AI transition. By integrating real-world experience with skill development, apprenticeships can effectively prepare the workforce for emerging roles in an AI-driven economy. Despite the potential benefits, only one country – Costa Rica – reports actively utilising apprenticeship incentives to support workforce development tailored to the specific needs of the AI economy, highlighting a significant opportunity for other countries to adopt similar strategies to ensure their labour forces are equipped for the AI transition.
Promotion of private‑public collaborations
Promoting collaborations between the private and public sectors can assist companies in adopting new AI technologies and providing effective training to their workforce, enabling them to fully capitalise on the opportunities brought forth by this innovative technology. By offering consulting and expertise services, governments can support companies to navigate the complexities of AI adoption, ensuring they harness the full potential of these transformative technologies while mitigating risks. This support can take many forms, including guidance on implementing AI training for staff, ensuring regulatory compliance, and providing access to expert resources from AI-focused government entities. Moreover, governments can facilitate collaboration between businesses and research institutions, fostering innovation and knowledge exchange in the AI domain. Only five out of 21 countries currently promote such collaborations.
As part of Germany’s AI Strategy, mobile and stationary “AI Studios” have been set up throughout the country to inform a broad range of sectors and companies, specifically SMEs, about AI in a user-friendly manner. The AI Studios aim at explaining AI technologies and their implications, empowering workers and social partners to actively participate in collaboratively designing the use of AI in the workplace. The project is being developed by two German research institutes (the Institute of Human Factors and Technology Management IAT, and the Fraunhofer Institute for Industrial Engineering) at the University of Stuttgart and funded by the Ministry of Labour and Social Affairs with around EUR 4.1 million. Similarly, in Austria, five National Digital Innovation Hubs and four European Digital Innovation Hubs were set up around the country between 2020‑22. The hubs provide support and expertise to SMEs that wish to digitalise or incorporate AI in their work. The programme was created by the Ministry of Labour and Economy and the National Foundation for Research, Technology and Development. Hungary has also launched a project to enhance the effectiveness and revenue‑generating potential of local small and medium-sized enterprises (SMEs) by helping them with the adoption of AI technologies. The project aims to offer business development and technology consulting, training sessions, self-assessment tools and capacity development to SMEs keen on harnessing AI and data resources. The project, co-financed by the European Regional Development Fund, is designed to benefit a total of 150 SMEs.
Deutsche Telekom, one of Europe’s largest telecommunication companies with around 200 000 employees in 34 countries, maintains a comprehensive internal training strategy to keep up to date with global developments and evolving skill needs. While most training is offered in a decentralised manner by the service providers on the ground, around 20% of training opportunities are managed centrally by Telekom’s Corporate Learning and Development Division. In total, Telekom offered 47 005 courses via its global Learning Management System in 2023.
As the introduction and proliferation of AI is expected to affect its business operations, Telekom has developed dedicated courses and learning journeys to improve AI literacy and the efficient use of AI tools. While Telekom designs and implements general AI training courses internally, more specialised courses tailored to specific employee groups are outsourced to external partners and learning providers to deliver.
An AI Explorer Course was introduced in 2023 to provide an overview of AI essentials. It spans 5 weeks with a minimum of four hours of learning time per week. With over 80 000 employees electing to participate in the training, it is evident that AI is of significant interest among the workforce.
In monitoring the impact of training, a key performance indicator for Telekom is the proportion of Digital Experts within the company. The percentage rose to 22% in 2023 from 13% in 2020. To further increase the share, Telekom introduced so-called “Explorer Journeys” in 2022 – central programmes which address specific digital innovation topics. In 2022, approximately 5 200 employees enrolled in multi-week digital training programmes covering various topics, including artificial intelligence, big data, digital marketing and software development.
Source: Based on a telephone interview with representatives from Deutsche Telekom’s Learning and Development division.
Incentives for adult learning providers
Another way to incentivise the provision of AI-related training is to focus on the supply side of training by directly or indirectly supporting adult learning providers, such as vocational colleges, universities and private institutions. This section highlights efforts to support adult education institutions to: 1) develop or update new training programmes, 2) update occupational standards and qualification frameworks, 3) create practitioner networks, and 4) offer train the trainer programmes (see Table 4.5).
Source: 2023 OECD Policy Questionnaire: Adult Learning for Diffusion of AI.
Updating or developing training programmes
While not yet common practice, public funding has been made available in some countries (nine out of 21) to incentivise education and training institutions to develop new training programmes delivering AI-related content. In Poland, with financial support from the EU, the Ministry of Family and Social Policy is funding the provision of second-cycle studies in the field of AI, machine learning and cybersecurity. In France, as part of the national AI strategy, a call for expressions of interest with a budget of EUR 2.5 billion has been created to financially support initial and continuing education providers (including universities, schools, apprentice training centres and private training organisations) in their development of new training programmes related to AI, as well as the training of the first cohort of students. In Croatia, the Ministry of Labour, Pension System, Family and Social Policy is currently overseeing the development of post-graduate training programmes focusing on digital skills, including AI skills, together with the Faculty of Electrical Engineering and Computing at the University of Zagreb.
A pre‑existing grant programme in Norway, the Tripartite Industry Program, is likely to be used to support the development of AI-related training. Within the framework of the Competence Programme devised by the Ministry of Education, grants are given to education providers with the aim of fostering the development of new continuing and further education opportunities across five designated industries for the year of 2025. The selected number of designated industries varies from year to year. Eligible recipients encompass accredited universities, publicly supported public and private colleges, public and private vocational schools with public support, adult learning associations, as well as other private education providers. Allocation of grants is conducted through a call for expressions of interest, with the Norwegian Directorate of Higher Education and Skills and the social partners placing particular emphasis on further education initiatives that bolster digital competence, ICT security, and competencies aligned with the green transition and sustainability.
Another approach is to update curricula of existing training programmes to deliver more AI-related content. For instance, Hungary introduced new sections linked to AI in the training and outcome requirements for several majors. More countries can consider updating existing training curricula to ensure programmes remain relevant for the new world of work.
Updating occupational standards and national qualification frameworks
Governments can also support the regular updating of qualification frameworks and occupational standards in order to ensure they reflect up-to-date skills requirements, including those related to AI. In Korea, this is achieved by the regular revision of the National Competency Standards (NCS), which reflect current industry trends and advancements, particularly in emerging technologies including AI. The Industrial Skills Council is actively engaged in the development of the NCS, ensuring that it’s shaped by the insights and perspectives of companies and industries.
In Australia, ensuring that qualification frameworks and occupational standards remain robust and responsive to emerging trends in AI is one of the priorities of a new inquiry initiated by the House of Representatives Standing Committee on Employment, Education and Training into the use of AI within the Australian education system. This inquiry aims to explore the issues and opportunities associated with generative AI and its potential impacts on early childhood education, schools and higher education sectors. In parallel in Australia, the Tertiary Education Quality and Standards Agency (TEQSA), as the national regulator of higher education, collaborates with stakeholders and experts to navigate the challenges posed by AI. TEQSA have released comprehensive resources on its website to assist the higher education sector in adapting to the implications of AI, covering areas such as policy development, academic integrity, and assessment design considerations. This helps providers update their programmes with relevant AI content.
Creating practitioner networks
Establishing or strengthening networks to foster better sharing of best practices and innovations in the development and implementation of AI-focused training programmes can help education and training institutions enhance the quality and relevance of educational offerings in response to evolving technological demands. However, among the 21 countries that responded to the policy questionnaire, only one – Austria – reported actively promoting practitioner networks that facilitate this exchange, revealing a potential area for growth.
Train the trainer programmes
Trainers and teachers need to be equipped with relevant knowledge and competences to effectively deliver AI-related training. Four of the 21 countries that responded to the policy questionnaire reported offering train the trainer programmes designed to enhance trainers’ capabilities in delivering AI-focused education.
In Costa Rica, for instance, the Ministry of Public Education is offering courses aimed at enhancing teachers’ skills and knowledge in the field of AI. These courses cover topics such as Mathematical and Statistics Probability Foundations for AI, AI Database, AI for Youth, and AI Programming. With the rise of AI affecting many occupations, more train the trainer programmes may be necessary in the future.
An analysis of training provision in Australia, Germany, Singapore and the United States sheds light on the quantity and delivery of AI-related training. For the purposes of this analysis, courses were identified as having AI elements if a keyword in an AI concept list (see Annex B) was found in the course title or description. Based on this analysis, the share of courses that include AI elements ranges from 5.5% in Singapore (2 613 courses), 2.5% in Germany (82 684 courses), to only 0.6% in the United States course catalogue (189 courses) and only 0.3% in Australia (44 courses). Below is a sample of the course titles that were identified as AI-related in each of the four countries (Table 4.6) and an illustration of one AI-related course description for each country, with the relevant AI keywords in bold (Table 4.7).
These estimates likely represent a lower bound estimate for the true share of AI-related courses in these countries: for the AI content to be mentioned in the course title or description, it needs to be quite substantial. It is likely that some courses have started to incorporate elements of AI-related training into their syllabus but do not mention this in the title or description. Moreover, training delivered by the non-formal sector is minimally covered in this analysis, while training delivered directly by employers is not captured at all. With the fast pace of development of AI, it is likely that employers or private training providers are developing training materials at a faster pace than the formal adult learning sector, thereby satisfying some demand.
Source: Australia’s NCVER vocational training database; Germany: IWWB course catalogue; Singapore: Skills Future Singapore course database; United States: Credential Engine database.
Note: Keywords highlighted in bold. Both course description and course objective provided for Germany and Singapore.
Source: Australia’s NCVER vocational training database; Germany – IWWB course catalogue; Singapore – Skills Future Singapore course database; United States – Credential Engine database.
This section describes the characteristics of AI-related training provision. As sample sizes for AI-related courses are small in Australia and the United States, caution is warranted when interpreting the descriptive findings and general extrapolations can only confidently be made using data for Singapore, given the relatively higher share of AI courses here. This data tell us that in general, AI-related courses are slightly more likely to be offered online (Figure 4.1) and are more targeted towards developing advanced skills when compared with the average adult learning course. This suggests that at the moment, most AI-related training is focused on training for AI professionals. This is despite the existence of broader terms such as “artificial intelligence” – which would capture general AI literacy courses – in our keyword list. Going forward, there is room to expand delivery of courses covering more general AI literacy.
As noted above, in Singapore, 40% of AI-related courses are offered online, slightly higher than the share for all training courses (31%). AI-courses also last on average about 89 hours, shorter than the average course (124 hours). This suggests that there is some flexibility in delivery mode and duration. Furthermore, a post-secondary degree or diploma was more likely to be a pre‑requisite for AI-related training than average training (29% versus 19% for all courses), which suggests that AI-related training may be more advanced or targeted towards higher skilled adults. At the same time, 25% of AI-related courses do not have a minimum entry requirement or only require primary school level education, slightly higher than the rate for the entire sample (19%), indicating reassuringly that there are also AI-related courses on offer for the broader population.
Similar to Singapore, AI-related training in Australia tends to be more advanced than average training, with AI courses more likely to be offered as part of a diploma or advanced diploma, and by universities. Consistent with the more advanced nature of AI-related training in Australia, courses also tend to last longer, with an average duration of 65 hours (compared to 38 hours for all training). Moreover, only 22% of courses are short (compared to 75% amongst the entire sample). The vast majority of AI-related courses (78%) are also classroom-based – as opposed to 42% for the entire sample – suggesting AI training in Australia may be less flexible in delivery mode and may require more practical or hands-on training.
Similar to Singapore, AI-related training in the United States is more flexible with 27% of AI-related courses adopting a blended delivery mode (compared to 18% for the entire sample) and 61% utilising a self-paced format (compared to 49% for the entire sample). The share of training offered completely online is the same for AI and all courses (41% for each). Furthermore, AI-related training is much shorter than the average education programme, requiring on average 16 weeks of student contact, compared to 53 weeks for the entire sample. In-line with this finding, AI courses are less expensive than the average course. However, no AI-related courses have a duration of less than 50 contact hours (compared to 0.55% of all courses), rendering these courses still relatively long. AI-related training is also more likely than the average to be offered by not-for-profit providers (68% for AI courses versus 39% for all courses), and less likely to be offered by public institutions like universities (30% versus 57%).
Unlike Singapore and the United States, but similar to Australia, the majority of AI-related courses in Germany require in-person presence (69%). Although this is slightly less than the share of all courses that are offered in presence (81%), it still makes the offer of AI-related courses quite inflexible. Only a small share of courses are web- or computer-based courses (8%) or are offered as distance learning courses (2%). This data therefore suggest that AI courses could be delivered in more flexible ways in order to incentivise greater uptake in the future.
Additionally, the data from Australia allows some insights into the characteristics of students enrolling in AI-related training. In 2022, just over 1 500 students were enrolled in at least one AI-related subject. These students tended to be younger on average than students enrolled in other courses (mean age 28 versus 35). The part-time enrolment rate amongst AI learners was also significantly lower than for all learners (43% versus 94%), with many more unemployed or out of the labour force while studying full-time. The student pool of AI learners is less gender-diverse with only 30% identifying as female (compared to a more equal balance of 47% in the entire sample). This reflects the broader literature which finds much lower rates of women enrolled in education with technical specialisations such as information and communications technology and science, technology, engineering and mathematics (STEM) (World Bank, 2020[8]; World Economic Forum, 2022[9]). Despite this lack of gender diversity, AI courses do appear to represent greater cultural and racial diversity, with more international students enrolled in AI courses (15% versus 4%).
Percentage point difference in share of courses by delivery method (AI courses versus all courses)
Note: For the United States, the percentage point difference in share of courses for online mode of delivery is zero. Information on hybrid mode of delivery unavailable for Germany and Singapore. Given underlying differences in data, cross-country comparisons are not advised.
Source: Australia’s NCVER vocational training database; Germany: IWWB course catalogue; Singapore: SkillsFuture Singapore course database; United States: Credential Engine database.
Policy discussions about how to prepare workforces for the diffusion of AI within workplaces often focus on this transition in a vacuum, independent of other transitions like the green transition. But of course, labour markets are facing various transitions all at once and a holistic approach to upskilling and reskilling which considers both these transitions jointly may be warranted in some contexts, for instance in sectors or jobs that require the use of AI technologies to support the green transition.
The analysis of training catalogues showed that training courses that are both green- and AI-related at the same time are rare: only 0.02% of total courses in Australia (3 courses), 0.12% in Germany (4 077 courses), 0.04% in Singapore (20 courses), and 0.02% in the United States (6 courses) could be considered both green- and AI-related. Table 4.8 provides examples of course titles for courses that are identified as being both green- and AI-related in the analysis. Table 4.9 provides an illustration of one course description for each country, with the relevant green and AI keywords in bold. Whilst the share of courses tackling the twin transition is very low across the countries studied, it is worth noting that a joint approach may only be efficient in certain contexts and therefore independently expanding initiatives for each transition should remain the focus in most countries.
Source: Australia’s NCVER vocational training database; Germany: IWWB course catalogue; Singapore: Skills Future Singapore course database; United States: Credential Engine database.
Note: Keywords highlighted in bold. Both course description and course objective provided for Germany and Singapore.
Source: Australia’s NCVER vocational training database; Germany: IWWB course catalogue; Singapore: Skills Future Singapore course database; United States: Credential Engine database.
Nonetheless, a handful of countries have taken a joint approach to upskilling and retraining for the twin transition. Both Croatia and France for instance offer financial incentives to support the development of skills related to digital and green content. In Croatia, vouchers to subsidise the cost of adult learning are available specifically for training programmes that deliver either digital or green skills. In France, two policy initiatives focus on jointly meeting the needs of the digital and green transitions. The Human Resources Consultancy Service (PCRH – Prestation de conseil en ressources humaines) supports small and medium‑sized businesses in their HR challenges in the face of the ecological and digital transitions. As part of France 2030, a wide‑reaching plan to invest in industrial competitiveness and the technologies of the future, the Ministry of Labor launched a call for expressions of interest in “Skills and professions of the future” (AMI-CMA). Successful proposals receive funding for skills assessments and new training programmes for the acquisition of skills in future jobs linked to the green and digital transitions.
In Hungary and Greece, policy initiatives consider the green and digital transitions together in both the content and delivery of learning. Hungary’s Highly Qualified, Competitive Workforce plan, part of the Hungarian Recovery and Resilience Plan, contributes to the modernisation of the vocational and higher education systems. It addresses the challenges of the green and digital transition by implementing energy efficiency renovation and digital equipment solutions in buildings in higher and vocational education institutions. Greece’s public employment service (DYPA) runs Apprenticeship Vocational Schools (DYPA EPAS), which implement the dual VET system in fields of study that vary each year, and in 2023‑24 will focus on the information society (“Smart” Electronic Devices and Facilities Technician, Support of Information System Technician, and Cybersecurity). The upgraded curricula and laboratories have taken environmental protection and green economy considerations into account and there are plans for integrating digital technologies in the educational process, especially through “virtual classrooms”.
References
[6] Hamari, J., J. Koivisto and H. Sarsa (2014), Does gamification work? - A literature review of empirical studies on gamification, IEEE Computer Society, https://doi.org/10.1109/HICSS.2014.377.
[3] Long, D. and B. Magerko (2020), “What is AI Literacy? Competencies and Design Considerations”, ACM Digital Library, Vol. CHI '20, pp. 1-16.
[4] Ng, D. et al. (2021), “AI Literacy: Definition, Teaching, Evaluation and Ethical Issues”, Proceedings of the Association for Information Science and Technology, Vol. 58/1, pp. 504-509, https://doi.org/10.1002/pra2.487.
[2] OECD (2023), OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en.
[7] OECD (2021), Training in Enterprises: New Evidence from 100 Case Studies, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/7d63d210-en.
[5] OECD (2019), OECD Skills Outlook 2019: Thriving in a Digital World, OECD Publishing, Paris, https://doi.org/10.1787/df80bc12-en.
[1] 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.
[8] World Bank (2020), Share of graduates by field, female (%) - World Bank Gender Data Portal, https://genderdata.worldbank.org/indicators/se-ter-grad-fe-zs/?fieldOfStudy=Science%2C%20Technology%2C%20Engineering%20and%20Mathematics%20%28STEM%29&view=bar (accessed on 7 June 2023).
[9] World Economic Forum (2022), Global Gender Gap Report 2022, World Economic Forum, Geneva, https://www3.weforum.org/docs/WEF_GGGR_2022.pdf.
← 1. Gdańsk University of Technology, Poznan University of Technology, Wrocław University of Science and Technology, Adam Mickiewicz University in Poznań and the University of Warsaw.