This chapter provides an overview of how countries develop curricula and qualifications in vocational education and training (VET). It maps the common processes involved in translating occupational requirements into learning outcomes, curriculum content and certification criteria, and highlights the complex interplay among national VET authorities, sectoral bodies, employers, trade unions and VET providers in this development process. Drawing on OECD survey evidence, the chapter identifies key challenges in current VET development processes and highlights areas where AI could help improve efficiency, responsiveness and alignment with labour market needs.
Developing Vocational Education and Training with Artificial Intelligence
Abstract
The transformative potential of AI in VET development: Challenges it can help address
Unique context of VET development: Vocational education and training (VET) curriculum and qualification development across OECD and European Union (EU) countries operates within complex, multi-layered governance arrangements (e.g. cross-ministries, cross-sectors, public-private and national-local) that translate labour market needs into occupational standards, qualifications and curricula. While the development process widely varies across countries – from centralised, legally regulated systems to decentralised, provider-led models – these three components are closely interlinked and typically developed through structured, stakeholder-driven processes. Industry actors often lead occupational standards and qualification design, while education authorities and VET providers play a central role in translating these into curricula aligned with national qualification frameworks and evolving labour market demand.
Challenges in VET development: While such multi-stakeholder and multi-level arrangements strengthen relevance, rigour and legitimacy, they are lengthy, resource‑intensive and co‑ordination-heavy processes – still with data gaps and uneven stakeholder representation – that struggle to keep pace with rapid change driven by digitalisation, green transitions, emerging technologies as well as lifelong learning demands. These pressures are compounded by the need to balance diverse stakeholder interests, manage large and increasing volumes of standards and curricula, and interpret fragmented and fast‑changing labour market information – draws on multiple sources, including legal frameworks, labour market intelligence, skills anticipation and expert input. This underscores the growing demand for more efficient analytical and development tools to support competency mapping and ensure relevance.
Using AI to improve VET development: Within this context, AI emerges as a complementary tool that can enhance efficiency, relevance, evidence synthesis and analytical capacity without replacing human judgement or collective decision making. Given current AI capabilities and stakeholder acceptance, its most plausible contributions lie in: (i) supporting stakeholder consultation through structured synthesis, (ii) streamlining drafting and compliance checks, and (iii) improving labour market intelligence by integrating and analysing large‑scale, heterogeneous data sources. Used within robust governance, quality assurance and ethical frameworks, AI can help VET systems move towards more agile, timely and evidence‑informed development while preserving legitimacy and educational and training quality.
To understand how artificial intelligence (AI) can support the development of vocational education and training (VET) curricula and qualifications (thereafter, VET development), it is necessary to examine how these are currently developed, interconnected and aligned with labour market demand. Across OECD and EU countries, VET development is shaped by different operational and procedural arrangements, stakeholder roles and sources of input. Drawing from the OECD surveys conducted as part of this study, this chapter ultimately builds the informative basis about VET development for Chapter 3, which explores how AI can be effectively leveraged in the VET development process with concrete country examples.
How is the development of VET curricula and qualifications organised?
VET curricula are structured education and training plans that specify programme content, delivery processes and pedagogical approaches (see Box 1.2, in Chapter 1). They typically serve as the principal pathways to VET qualifications, which are often based on occupational standards and developed with strong involvement of social partners.
Countries organise VET development in different ways. In countries such as Croatia, Germany, Lithuania, Mexico and Switzerland, dedicated legal and regulatory frameworks set the rules and procedures for developing national occupational standards, qualifications and, in many cases, the associated curricula. In other countries such as England (United Kingdom), Ireland and Korea, accredited VET providers or awarding bodies enjoy varying degrees of autonomy in designing curricula or qualifications – usually within centrally defined structures, such as employer‑developed occupational standards (England), programme validation and awards standards frameworks (Ireland), or national competency standards (Korea). Some systems combine both approaches: for example, in Estonia, Finland and the Netherlands, qualifications are centrally defined, while VET providers retain autonomy to develop curricula or locally tailored units within national qualification requirements (OECD, 2025[1]).
The development of occupational standards, VET qualifications and curricula is inter-linked
Occupational standards or qualifications define overall occupational structures, competency requirements and learning outcomes, while curricula translate requirements into teachable and assessable programmes. In this sense, the development of occupational standards, qualifications and curricula is inter-linked, and involves a process of co-development between industry and education (Figure 2.1).
Occupational standards – also called professional or sectoral standards – define the skills, knowledge, abilities and attitudes needed for effective performance in specific occupations, often serving as national or sectoral competency frameworks that guide workforce development, VET qualification and curricula design (OECD, 2024[2]).
Qualifications formalise these standards into certification frameworks, typically mapped to National Qualifications Frameworks (NQFs) (Gasskov, 2018[3]) – structured learning and assessment requirements for certification, thus often include criteria for examination, accreditation and certification.
Curricula translate qualification requirements into structured VET programmes, specifying educational goals and core competences, course content, teaching methodologies and assessment methods, and thus provide teaching and learning plans to deliver training and assess learners (Cedefop, 2015[4]). NQFs are the main tool for ensuring quality in the alignment between standards and qualifications to curriculum.
Note: This figure presents a simplified overview of VET curricula and qualifications, summarising national terminologies used to refer to the three elements. Each country also applies different processes and methodologies.
Source: Author’s elaboration based on OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
While closely connected, the degree of integration between standards, qualifications and curricula varies across countries (Table 2.1). Some systems follow sequential development processes (e.g. Croatia, Estonia, Lithuania and the Netherlands), while others integrate some of these processes more closely (e.g. Australia, England, Germany and Switzerland) or manage parallel but linked processes (e.g. Korea). This is particularly evident in apprenticeships and work-based programmes where occupational standards, training plans and assessments are often directly embedded in curricula, as in England, Switzerland and Germany. In England, apprenticeship content, training plans and end-point assessments are mapped directly to occupational standards, and the curriculum is set out in the apprenticeship standard and assessment plan. In Switzerland, VET ordinances integrate occupational profiles, training plans and qualification procedures, forming the backbone of dual VET. In Germany, training regulations and framework curricula guide both in-company training and school-based VET, ensuring harmonised and standardised content and assessment across regions and companies (Table 2.1).
Note: Case study countries are highlighted in blue.
Source: OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
Development process of VET curricula and qualifications: Structured and stakeholder-driven
Across countries, VET development follows a structured, multi-phase and stakeholder-driven process, involving labour market analysis, occupational profiling defining competencies and specifying qualification requirements, pedagogical design and assessment planning. Stakeholder engagement – particularly with employers, sector bodies, education experts and labour market intelligence units – is central throughout. While revision cycles vary across countries and sectors, most VET systems include periodic reviews (often every 3‑5 years), complemented by ad hoc updates to respond to technological or economic change (OECD, 2025[1]).
VET development typically begins with comprehensive input gathering through labour market analysis, employer surveys and stakeholder insights – and increasingly supported by AI techniques such as large language models (LLM) and natural language processing (NLP). These inputs inform the drafting and refining of occupational standards, qualification profiles and curricula. Draft outputs are refined through consultation process, validated and approved within national regulatory and quality assurance frameworks, considering consistency across sectors and occupations. Governance arrangements often combine national co‑ordination with sectoral or provider‑level responsibility, balancing consistency with flexibility. Increasingly, VET systems are embracing flexible pathways, micro-credentials and recognition of prior learning to support lifelong learning and inclusion, with a growing need to respond to digitalisation, AI, green skills and adaptability to economic shifts.
Which stakeholders are involved in VET curriculum and qualification development?
VET development is driven by a wide range of stakeholders who collectively ensure that VET programmes remain relevant, high quality and responsive to labour market and educational needs. Governments and policymakers establish national strategies and regulatory frameworks, while industry representatives often lead occupational standards and qualification design; and VET providers and teachers contribute to curriculum development and pedagogical alignment. In several countries – including Belgium, Estonia, Finland, the Netherlands, Portugal and the Slovak Republic – providers enjoy autonomy to design curricula within national qualification requirements, tailoring content to regional labour market needs, employer partnerships and institutional specialisation and resources. At the same time, national co‑ordination bodies or qualification authorities (e.g. BIBB, QQI, SERI, HRDK and SBB) play a critical role in ensuring coherence, quality and alignment with national and European/international frameworks.
VET development is also multi‑layered, spanning national, regional and local levels, and often involving multiple ministries and sector‑specific authorities. These arrangements support relevance and adaptability but require intensive co‑ordination, contributing to longer development cycles.
Multiple stakeholders are involved in VET curriculum and qualification development
Multiple stakeholders are involved in VET development (Table 2.2). Typically, industry representatives lead the development of occupational standards and qualifications, while education representatives – including VET providers, curriculum specialists and teachers – tend to lead curriculum development. Both domains, however, influence each other significantly. Government agencies, such as VET authorities and qualification regulators, play a central role in co‑ordinating processes, balancing perspectives and ensuring compliance. Additional support is often provided by mandated experts or specialised bodies – for example, universities, research institutions or consultancy organisations – who undertake technical drafting, conduct analyses or facilitate the development process, often compensating for the limited capacity of stakeholders whose main responsibilities lie outside curriculum and qualification development.
Note: Case study countries are highlighted in blue.
Source: OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
Multi-layer development and implementation of VET curriculum and qualifications
VET development is not only multi-stakeholder, but also multi-layered process. Many countries develop VET curricula and qualifications at different levels – national, regional, local or institutional levels. Even in countries where VET providers do not play a significant role in the development (e.g. England [United Kingdom], Germany and Switzerland), VET schools and teachers often have autonomy to adapt content within the parameters of national or regional regulations.
Local VET curriculum and qualification development offers flexibility and speed: In several VET systems, responsibilities for VET curriculum and qualification development are distributed across national, regional and local levels.
In Finland, alongside nationally defined qualification units, VET providers or local businesses may design local units tailored to regional labour market needs or choose to develop local curricula based on national qualification requirements. These local units – limited to 15 competence points – do not require national approval, provided they remain within the provider’s licence and comply with national competence requirements. They can be elevated to national units if they demonstrate broader relevance (OECD, 2025[7]).
In Canada, provinces and territories develop and update VET curricula and qualifications in collaboration with employers, unions and VET institutions through surveys, focus groups and working groups. Consistency across more than 50 skilled trades is supported by the federally administered Red Seal Program. Red Seal Occupational Standards (RSOS), which outline Red Seal curricula and qualifications, are developed by the Canadian Council of Directors of Apprenticeship (CCDA) – which includes federal, provincial, and territorial apprenticeship authorities – and guide both federal and provincial and territorial curricula (apprenticeship training) and certification examinations. VET institutions adjust curricula accordingly, and Red Seal exam blueprints are revised to reflect new standards (Red Seal Program, 2024[8]).
Private qualifications and curricula promote flexibility and speed: In some countries, private qualifications and curricula exist to give flexibility.
In Korea, national qualifications are expected to align with the National Competency Standards (NCS), while private qualifications are developed by private entities offering flexibility and responsiveness to emerging or niche skill demands.
Qualifications and curricula developed by multiple ministries and sector-specific standards: VET qualifications and curricula are often developed through collaboration across multiple ministries, administrative branches and sectoral authorities to ensure alignment with sector-specific legislation and standards.
In Finland, the Ministry of Education and Culture (MoEC) oversees overall VET governance, including licensing education providers and structuring qualifications. However, qualifications in sectors such as social welfare and healthcare, transport, real estate, construction and natural resources are governed by statutes and regulations specific to those sectors, under the purview of relevant sectoral ministries. The Finnish National Agency for Education (EDUFI) prepares national qualification requirements, while sectoral ministries ensure that qualifications meet legal and professional standards relevant to their domains. This multi‑authority approach ensures that qualifications are pedagogically sound, legally compliant and responsive to labour market needs.
Similarly, in Korea, nationally regulated qualifications are administered by 29 ministries under 94 sector-specific laws. For instance, sector-specific bodies such as the Korea Health Personnel Licensing Examination Institute (under the Ministry of Health and Welfare) and the Ministry of Employment and Labour oversee healthcare and technical qualifications respectively.
In Lithuania, sectoral standard groups and the Qualifications and VET Development Centre co‑ordinate the development of sectoral standards and qualifications, with input from VET providers and employer associations.
These examples illustrate the multi-level, cross-sectoral, multi‑agency and public-private nature of VET development and governance, which helps ensure that qualifications remain relevant, agile and responsive to evolving labour market needs, but at the same time contributes to lengthy, complex and demanding processes that require intensive co‑ordination.
What are the key sources of inputs for VET curricula and qualifications?
VET development in many countries draw on multiple inputs, including national frameworks and standards to skill intelligence that includes labour market data and VET stakeholder perspectives. The OECD Surveys conducted as part of this study (OECD, 2025[1]) shows that national frameworks, inputs from sector, industry and providers as well as labour market data are the primary sources for developing VET curricula and qualifications: 78% of VET providers and 64% of industry partners use national frameworks, such as legislations, occupational/competency standards, national qualification frameworks, as primary source. 73% of VET providers and 93% of industry partners reported using sector and industry inputs as their primary sources, and 59% of providers and 51% of industry partners reported using labour market data as their primary sources. 63% of providers use their own inputs (Figure 2.2).
Share of survey respondents
Source: OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
Key source of input 1: Legal frameworks defining the development of VET curriculum and professional standards/qualifications
VET development in many countries is governed by a combination of national legislation, ministerial decrees and sectoral regulations. VET laws typically designate the key agency that oversees the relevance and quality of VET, ensuring that VET curricula and qualifications meet national standards and rules.
For example, in Switzerland, the Vocational and Professional Education and Training Act (VPETA) and associated ordinances issued by the State Secretariat for Education, Research and Innovation (SERI) provide the legal foundation for the development, implementation and periodic review of VET programmes and qualifications. These ordinances, developed by professional organisations and approved by the SERI, define the competencies to be acquired, the rules for implementation, and the procedures for qualification.
In Croatia, the process is defined by the Vocational Education Act, the Act on the Croatian Qualifications Framework (CROQF), and a series of methodologies and guidelines issued by the Agency for Vocational Education and Training and Adult Education (AVETAE or ASOO). These legal instruments stipulate that VET curricula must be competency- and learning-outcome‑based, grounded in occupational and qualification standards, and aligned with the CROQF.
England provides another example, where the development of VET curricula and qualifications is governed by a combination of Department for Education (DfE) policy frameworks, SkillsEngland/IfATE (Institute for Apprenticeships and Technical Education, former SkillsEngland) guidance, and regulatory oversight by Ofqual. While their occupational standards are developed by employers, the legal and policy framework ensures that qualifications and curricula developed based on those standards are aligned with labour market needs and national skills priorities, and that they undergo rigorous approval and quality assurance processes.
Laws and regulations that countries must adhere when developing VET curricula and qualifications
Note: Case study countries are highlighted in blue.
Source: OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
Key source of input 2: Skills anticipation, skills intelligence and labour market data
VET development is informed by skills needs assessment and anticipation or skills intelligence. Skills anticipation (or skill needs assessment and anticipation) refers to the strategic evaluation of current and future skill needs in the labour market, using consistent and systematic methods. It involves identifying skill gaps and shortages, and evaluating the capacity of qualification systems and VET provisions to meet demand. This process can take place at national, regional, local or sectoral levels and provides quantitative (e.g. employment and job vacancy trends by occupation or qualification) and qualitative (e.g. evolving skill profiles within jobs) insights into emerging skill needs and future gaps (Cedefop & UNESCO-UNEVOC, 2025[9]). Skills intelligence encompasses the identification, collection, analysis, synthesis and presentation of labour market and skills data. It supports the anticipation and forecasting of skill needs, addresses gaps and mismatches, and informs education and training provision. It offers insights into labour market profile and employment trends, occupational requirements, working conditions, pathways through education and employment (Cedefop & UNESCO-UNEVOC, 2025[9]).
There are several skills anticipation methods including skills forecasts, foresights and surveys as well as graduate tracking and sectoral analyses. Increasingly, big data and other AI tools are used to assess and anticipate skill and jobs developments, especially in the context of digital transformation and the green transition. Online job advertisement data provides (quasi) real-time insights into employer demands, complementing traditional tools such as stakeholder workshops and labour market statistics. These trends help map skills shortages – both short, medium and long term – and alleviate them, by guiding adjustments in VET curricula and qualifications (Cedefop & UNESCO-UNEVOC, 2025[9]).
Skills anticipation, skills intelligence or labour market data are used at the beginning of the VET development process, as input. Croatia and Lithuania are required to conduct a labour market needs analysis and employer surveys to identify current and future skills gaps. In Ireland, VET providers consult labour market data, employer feedback and national skills reports (e.g. from SOLAS and QQI) to design and update VET curricula and programmes. Similarly, in Switzerland, professional organisations conduct regular reviews of economic and technological trends, using labour market intelligence to update VET Ordinances and framework curricula. These approaches ensure that VET curricula are not only responsive to current employer needs but also anticipate future skills requirements, making use of both quantitative data (e.g. employment statistics, job vacancy analyses) and qualitative intelligence (e.g. expert panels, focus groups, and sectoral working groups) to guide curriculum development and qualification standards (OECD, 2026[10]).
In France, the Observatoires Prospectifs des Métiers et des Qualifications (OPMQ; French occupation and skills observatories) conduct sector-specific foresight studies to anticipate future skill needs, which directly inform the design and revision of training programmes and occupational standards. Their insights support policymakers, training providers, and social partners in aligning VET offerings with evolving labour market demands (OECD, 2019[11]). In Greece, the Central Council for VET (KSEEK), together with regional councils (SSPAE) is responsible for monitoring the labour market changes and qualification needs.
Key source of input 3: Input and output repository
Across many countries, VET development increasingly relies on robust input and output repositories – databases and platforms that capture labour market trends, skills registers, occupational standards, educational records, curricula and micro-credentials. When augmented by AI, these systems can automate the analysis of large‑scale data, identify emerging skills and generate actionable insights for curriculum design and policy decisions. By integrating and leveraging such repositories, countries can ensure that VET programmes remain relevant, responsive, and aligned with evolving workforce needs in a rapidly changing world.
Estonia is significantly advanced in this area. Estonian labour market and skills forecasting system (OSKA) provides labour market analysis and relevant input for VET development, while Skills Register linked to occupational standards and qualifications (via Occupation Register) and supports the comparison and validation of learning outcomes, certificates and diplomas across educational levels. Estonian Education Information System (EHIS) covering education institutions, registered curricula, their learning outcomes and student records. Estonia is also currently developing a registry for micro-credentials. Meanwhile, VET teachers in Estonia use social media groups and informal networks to share resources including prompts for AI use (e.g. Estonia), and these platforms could be also leveraged by AI to analyse emerging trends, support prompt engineering and facilitate peer learning (OECD, 2025[1]).
In Belgium (Flanders), Qualifications and curricula are made available through the Kwalificatiedatabank (Qualification Database). Occupational standards are included in the “Competent” database. The content of the database is managed and updated by the Flemish Public Employment Service (VDAB), while the database is a joint product of the Belgian public employment services Actiris, ADG, Bruxelles Formation, Forem and VDAB. Standards are periodically reviewed to ensure they remain relevant and responsive to changing labour market needs. Updates are informed by analyses of labour market data, including online job vacancies (OECD, 2024[2]).
In France, Occupational standards are organised through the ROME system (Répertoire Opérationnel des Métiers et des Emplois), which helps match job offers with job seekers, and is managed by the French Public Employment Service, France Travail. The recent transition to ROME 4.0 marks a significant update, with a shift to a skills-based approach to meet evolving labour market demands. A dedicated online platform, where sectoral funds, social partners, training providers and other key actors can suggest updates, review and provide feedback on draft profiles prepared by France Travail, has been developed to facilitate stakeholders’ involvement (OECD, 2024[2]).
In addition, there are international resources. For example, pan-European projects like the European Software Skills Alliance and Digital Europe enables the participating countries to contribute to and benefit from shared databases and frameworks for software and AI skills, which can be further enhanced by AI for benchmarking, automated curriculum generation, and best practice exchange (Box 2.1).
The European Software Skills Alliance (ESSA) and Digital Europe projects include participation from a wide range of European countries. For ESSA, participating countries include Belgium, Estonia, Germany, Greece, Hungary, Italy, Ireland, France, the Netherlands, Poland, Slovenia, Spain, Luxembourg and the United Kingdom. The Digital Europe Programme is even broader, involving all EU member states as well as EFTA/EEA countries (Iceland, Norway, Liechtenstein), and several candidate and associated countries such as Albania, Bosnia and Herzegovina, Kosovo, Moldova, Montenegro, North Macedonia, Serbia, Türkiye, Ukraine and Switzerland (with some limitations on objectives) (Digital Europe Programme, 2025[12]). These alliances and programmes foster cross-country collaboration on digital skills, software competencies and curriculum development, leveraging AI and shared resources for Europe‑wide impact.
VET systems operate in an environment of accelerating technological change, most recently with the emergence of AI. While digitalisation and online learning environments (Jeon, 2025[13]) as well as policy attempts such as modularisation, micro-credentials and competency-based frameworks have increased access, flexibility and relevance in VET (OECD, 2023[14]; OECD, 2023[15]; OECD, 2023[16]), formal curricula and qualifications often struggle to keep pace with evolving skill demands due to limited occupational or skills coverage, outdated content and lengthy revision cycles (Moein et al., 2024[17]). Established development processes – grounded in expert consultation, validation and compliance with national qualification frameworks – safeguard legitimacy and quality, but are resource‑intensive and difficult to adapt rapidly (OECD, 2025[1]). The resulting tension between stability and responsiveness raises a systemic question: how can VET systems shift from static, complex and periodic updates towards more agile, iterative and evidence‑informed approaches?
In this context, AI is increasingly viewed as a tool to support faster adaptation and improved relevance, provided that its use is embedded within robust governance and quality assurance arrangements. AI is expected to have efficiency gain and productivity improvement (Brynjolfsson, Li and Raymond, 2025[18]). Typical criteria for efficient occupational and training standard setting – such as responsiveness, timeliness, agility, user-friendliness and quality (OECD, 2024[2]) – are the areas where AI can help deliver improvements. Evidence from the OECD AI Capability Indicators suggest that current AI systems can perform intermediate‑level tasks in language processing, classification and pattern recognition (OECD, 2025[19]; OECD, 2023[20]), which are directly relevant to curriculum drafting, occupational profiling and large‑scale data analysis.
At the same time, AI systems face clear limitations: they may privilege easily codifiable or digitised skills and learning methods, potentially narrowing the breadth of curricula. Generative AI tools can introduce inaccuracies, cultural bias or opaque sourcing, complicating their integration into established validation procedures. Without involving VET stakeholders (e.g. in co-designing AI tools or AI-supported processes), AI-supported processes may shift or increase workload rather than reduce it, while raising concerns related to data governance and ethics. Moreover, current AI systems still lack the contextual judgement, robust reasoning, dynamic learning and social intelligence (OECD, 2025[19]; OECD, 2023[20]), which are required for VET development processes. Addressing these risks requires interdisciplinary collaboration, ethical governance, professional development and monitoring frameworks to ensure AI supports, rather than disrupts, educational equity, quality and effectiveness (see Chapter 3).
Within these constraints, AI is most plausibly positioned as an assisting tool in VET development to complement existing processes rather than replace them. Fully automated qualification design or curriculum revision remains beyond both current AI capabilities and stakeholder acceptance (OECD, 2025[1]). However, three key, interconnected areas illustrate where AI could help improve persistently challenging tasks in VET development (detailed discussion of each bullet point is under “Challenging task” section below and Chapter 3 explores country examples in a greater detail):
Assisting in balancing diverse stakeholder interests by synthesising processes and evidence (see Challenge 1 below): While AI cannot fully resolve the complexities of stakeholder co‑ordination, it can improve stakeholders’ preparedness, reduce and support their tasks such as streamline documentation, synthesise inputs and iterative consultation across multiple feedback rounds.
Streamlining lengthy and resource‑intensive processes (see Challenge 2 below): AI can help reduce time and resources throughout VET development processes that typically involve multiple stages of drafting, legal and regulatory procedures, compliance checks, validation and approval, often across several institutions and legal frameworks.
Improving labour market intelligence and relevance through integrating comprehensive insight and detecting gaps (see Challenge 3 below): AI can integrate and analyse large‑scale, heterogeneous labour market and education data to identify emerging skills, identify gaps between demand and existing provision, and support timelier, evidence‑informed updates to VET curricula and qualifications.
Challenging task in VET development 1: Balancing diverse stakeholder interests
VET development is inherently collaborative process, involving a diverse array of stakeholders. While this diversity ensures broad representation, thus legitimacy and relevance, it also introduces competing priorities: employers emphasise job-specific competencies, educators focus on pedagogical coherence and progression pathways, and regulators seek compliance with legal frameworks. Unequal capacity to engage – particularly among small and medium-sized enterprises (SMEs) – further complicates co‑ordination and slows decision making.
Co‑ordinating input and aligning goals across these diverse groups takes time and efforts and can lead to delays or conflicting priorities. These conflicting interests (see Figure 2.3), if not delicately balanced, may result in compromises that dilute the effectiveness of the curriculum and qualifications that are far from labour market demands, and delay the development, revision and reform cycles.
The OECD Surveys conducted as part of this study show that “balancing different stakeholder needs, interests and priorities” is among the most frequently cited challenges among VET policymakers (64%) (Figure 2.4, Panel A). Related challenges include “gathering comprehensive views and inputs” (56% of policymakers; 44% of industry representatives), “ensuring effective consultation, collaboration and co‑ordination across stakeholders” (56% of policymakers; 30% of industry representatives), and “balancing national standards with local flexibility” (reported by 47% of providers, 33% of industry representatives and 31% of policymakers) (Figure 2.4, Panel A). Balancing national and local needs is reported across a range of countries, where local autonomy can enhance responsiveness but also lead to inconsistent implementation that may undermine national coherence and comparability of VET curricula and qualifications – particularly during periods of structural transitions such as the shift to green skills (OECD, 2025[7]).
Source: Author’s elaboration, based on (OECD, 2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
While few stakeholders expect AI yet to fully address these issues (14‑24% reported “co‑ordination” as main motivations or objectives for using AI in VET development; Figure 2.4, Panel B), AI can support consensus-building indirectly by synthesising large volumes of stakeholder inputs, processes and evidence, and preparing stakeholders for working group meetings and advisory processes through concise briefs, enabling more informed discussions and reducing bottlenecks in decision making. AI has potential to offer new avenues for addressing co‑ordination challenges in VET development. Used transparently (i.e. clearly specifying when and how AI is used in VET development, for what purpose, on which data and with what limitations) (OECD, 2026[21]), such tools may improve the efficiency and inclusiveness of consultation without displacing collective decision making.
By facilitating structured input collection, synthesising diverse perspectives – some stakeholders noted AI’s perceived “objectivity” – and enabling scalable engagement, AI can support more inclusive and efficient consultation processes. Wider access to generative AI could help SMEs contribute more effectively to VET development processes by producing structured inputs for consultation and stakeholder engagement without requiring extensive technical capacity in translating company skills requirements into trainable curricula. AI-powered chatbots and data aggregation tools can summarise large volumes of employer feedback, learner surveys, and public consultations, making engagement more inclusive and efficient, fostering collaborative curriculum development and ensuring that VET qualifications reflect diverse stakeholder perspectives while reducing bottlenecks in consultation processes (OECD, 2025[22]).
Challenging task in VET development 2: Managing time and resources constraints
Developing high-quality VET curricula and qualifications requires substantial time, expertise, and administrative capacity. The scale of the task is also significant: countries manage hundreds of occupational standards, qualifications and curricula, many of which are extensive and highly detailed (Table 2.4). Development processes therefore take several years for new qualifications and nearly as long for updates, with initial VET processes in countries like Norway longer than continuing VET. Budget constraints could further limit capacity, scope and speed of development, stakeholder engagement and approval of new qualifications, while complex approval procedures hinder responsiveness to emerging industry needs – as reported in England.
The OECD Surveys conducted as part of this study indicate that a majority of stakeholders cite “lengthy procedures” (53‑61%) and “resource limitations” (56‑61%) as major challenges (Figure 2.4, Panel A). They Stakeholders primarily associate AI with “speed and scale” (64‑75%), reflecting its potential to process large datasets and stakeholder inputs, identifying and synthesising patterns, trends and perspectives that are difficult to detect manually, and reducing time and cost (Figure 2.4, Panel B). In this sense, AI is seen as a tool to compress time‑intensive analytical and co‑ordination phases that often delay VET development.
Stakeholders also expect that AI can “automate drafting” of curriculum components such as learning outcomes, competencies, module descriptions, assessment tasks and evaluation criteria within modules or learning outcome units (57‑75%, Figure 2.4, Panel B). AI’s benefits extend well beyond drafting alone. While few stakeholders (20‑30%) expect to resolve “regulatory constraints”, requirements and complexity (Figure 2.4, Panel A), AI can still assist with technical alignment by checking consistency and “compliance with legal and regulatory frameworks and standards” (7‑50% stakeholders reported, Figure 2.4, Panel A), reducing human error and time through automated checks. For example, AI can link learning outcomes to competency sets or credit points through automated checks (e.g. Croatia), and also accelerate consistent, accurate and efficient content generation at scale (e.g. England).
Emerging initiatives reflect this broader understanding of AI’s role. For example, the OECD’s Teaching Compass, complimented by the development of an AI curriculum analyser tool to support curriculum analysis and implementation has emphasised how generative AI can support curriculum transformation, including analysing curriculum content, supporting teachers’ curriculum implementation, and strengthening coherence between intended, implemented and learned curricula (see 2025 discussion from the OECD Global Forum on the Future of Education and Skills 2040). AI tools, particularly Natural Language Processing (NLP), can improve clarity, detect inconsistencies and support drafting while reducing manual workload (OECD, 2025[22]). Overall, AI’s potential to address time and resource constraints in VET development lies not only in automating drafting and reducing manual and repetitive work of experts, but also in restructuring how evidence is gathered, synthesised, translated into decisions and validated. This enables faster production of high-quality documents for VET systems, supporting more responsive updates to occupational standards, qualifications and curricula in VET.
Source: OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews. Korea from NCS Classification (NCS 분류) (n.d.[25]), NCS Classification, www.ncs.go.kr/th01/TH-102-001-02.scdo.
Challenging task in VET development 3: Anticipating and translating evolving labour‑market demand into VET curricula and qualifications
Ensuring the responsiveness of VET curricula and qualifications is a persistent challenge for VET policymakers, providers and industry partners. Established update cycles and consultation mechanisms often struggle to keep pace with rapid technological change, sectoral transformation and emerging skill requirements in a timely and sufficiently granular manner. While the results from the OECD Surveys conducted as part of this study (OECD, 2025[1]) suggest that relatively few stakeholders (16‑36%) perceive accessing and analysing labour market data as a major challenge, many interviewees report that available information remains fragmented, outdated or difficult to interpret, limiting its usefulness for forecasting skills needs and aligning curricula with emerging trends. These challenges are often linked not only to data availability, but also to limited interoperability across data sources: labour market, education and qualification data are often stored in disconnected systems, use different classifications, formats or structures, making it difficult to combine and analyse them systematically. Even when data are available, effective interpretation requires advanced analytical capacity and sector‑ and occupation‑specific expertise that is not consistently available across VET systems.
Within this context, AI can help address these gaps by integrating diverse data sources – from job vacancy postings and economic reports to employer surveys – and identifying patterns or mismatches that manual methods may overlook, including skills in high demand that are not covered by existing VET qualifications or programmes. This may enable earlier identification of emerging gaps, more evidence‑based decision making and timelier updates to occupational standards, qualifications and curricula, particularly in fast-changing sectors where traditional mechanisms may be too slow or insufficiently representative. The added value of AI lies not in replacing sector expertise, but in augmenting it with systematic, scalable evidence through the analysis of large volumes of heterogeneous data and the forecasting of future competency needs.
Recent research highlights the potential of Natural Language Processing (NLP) techniques to extract meaningful insights on skills acquisition from large volumes of textual content, reinforcing AI’s role in supporting curriculum development (Gonzalez-Gomez et al., 2024[26]). Increasingly, AI is used to analyse unstructured and large‑scale labour market data from job portals, employer databases and social media to anticipate emerging skill needs and identify new occupations, skill gaps and sectoral trends in near real time, informing curriculum design and updates (Cedefop & UNESCO-UNEVOC, 2025[9]). The OECD Surveys conducted as part of this study (OECD, 2025[1]) show that countries already apply AI techniques (machine learning and NLP) to skills forecasting and description, using big data analytics, real-time labour market data or AI-generated synthetic data – or linking curricula and occupational profiles with such data (e.g. Cedefop’s Skills OVATE and national initiatives in Australia, Canada, New Zealand, the United Kingdom and the United States). These applications demonstrate AI’s potential to improve responsiveness, efficiency and inclusiveness in VET development.
Moreover, stakeholders’ insight confirms that unlocking underutilised data sources is key to realising AI’s full potential in VET development, leveraging longitudinal student data, career guidance interactions and continuous feedback loops from industry and learners to inform agile updates and enable predictive analytics for skills forecasting (OECD, 2025[1]).
Percentage of respondents in each stakeholder group
Note: The number of respondents is in brackets.
Source: OECD (2025[1]), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews.
References
[18] Brynjolfsson, E., D. Li and L. Raymond (2025), “Generative AI at Work”, The Quarterly Journal of Economics, https://doi.org/10.1093/qje/qjae044.
[4] Cedefop (2015), The role of modularisation and unitisation in vocational education and training, Luxembourg: Publications Office. Cedefop working paper; No 26., https://doi.org/10.2801/38475.
[9] Cedefop & UNESCO-UNEVOC (2025), Meeting skill needs for the green transition: Skills anticipation and VET for a greener future, https://doi.org/10.2801/6833866.
[12] Digital Europe Programme (2025), List of Participating Countries in the Digital Europe Programme, https://ec.europa.eu/info/funding-tenders/opportunities/docs/2021-2027/digital/guidance/list-3rd-country-participation_digital_en.pdf.
[5] Finnish National Agency for Education (2025), Reform of a qualification requirement, https://www.oph.fi/fi/koulutus-ja-tutkinnot/uudistettavat-tutkinnon-perusteet.
[6] Finnish National Agency for Education (2023), Quality assurance of the qualification requirement development process at EDUFI, https://www.oph.fi/sites/default/files/documents/EDUFI%20QA%20of%20the%20qualification%20rquirements%20development%20process%20Finland%202023.pdf.
[3] Gasskov, V. (2018), Development of Occupational Standards, Qualifications and Skills Assessment, https://www.ilo.org/sites/default/files/wcmsp5/groups/public/@ed_emp/@ifp_skills/documents/projectdocumentation/wcms_645065.pdf.
[26] Gonzalez-Gomez, L. et al. (2024), Analyzing Natural Language Processing Techniques to Extract Meaningful Information on Skills Acquisition From Textual Content, https://doi.org/10.1109/ACCESS.2024.3465409.
[25] Human Resources Development Service of Korea (n.d.), NCS Classification (NCS 분류), https://www.ncs.go.kr/th01/TH-102-001-02.scdo.
[13] Jeon, S. (2025), “How can innovative technologies transform vocational education and training: Insights for Ukraine”, OECD Publishing, Paris, https://doi.org/10.1787/fb40f416-en.
[23] KRIVET (2025), 2024 Research Project for National Accreditation of Private Qualifications (민간자격 국가공인을 위한 조사연구 사업), https://www.krivet.re.kr/kor/sub.do?menuSn=12&pstNo=PB0000000422.
[17] Moein, M. et al. (2024), “Beyond Search Engines: Can Large Language Models Improve Curriculum Development?”, in Lecture Notes in Computer Science, Technology Enhanced Learning for Inclusive and Equitable Quality Education, Springer Nature Switzerland, Cham, https://doi.org/10.1007/978-3-031-72312-4_17.
[21] OECD (2026), “Five principles for the effective use of AI in vocational education and training development”, OECD Publishing, Paris, https://doi.org/10.1787/ee8c8e8e-en.
[10] OECD (2026), Ten Country Case Studies on VET Development and AI Use, OECD, Paris, https://www.oecd.org/content/dam/oecd/en/publications/support-materials/2026/06/developing-vocational-education-and-training-with-artificial-intelligence_e9f76b4e/Ten-country-case-studies-on-VET-development-and-AI-use.
[1] OECD (2025), Developing vocational qualifications and curricula with AI: Surveys and stakeholder interviews, Unpublished.
[22] OECD (2025), Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Functions, OECD Publishing, Paris, https://doi.org/10.1787/795de142-en.
[19] OECD (2025), Introducing the OECD AI Capability Indicators, OECD Publishing, Paris, https://doi.org/10.1787/be745f04-en.
[7] OECD (2025), Vocational Education and Training and the Green Transition in Finland, OECD Reviews of Vocational Education and Training, OECD Publishing, Paris, https://doi.org/10.1787/4d29a34a-en.
[2] OECD (2024), Agile Occupational and Training Standards for Responsive Skills Policies, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/bacb5e4a-en.
[14] OECD (2023), Building Future-Ready Vocational Education and Training Systems, OECD Reviews of Vocational Education and Training, OECD Publishing, Paris, https://doi.org/10.1787/28551a79-en.
[15] OECD (2023), “Micro-credentials for lifelong learning and employability: Uses and possibilities”, OECD Education Policy Perspectives, No. 66, OECD Publishing, Paris, https://doi.org/10.1787/9c4b7b68-en.
[20] OECD (2023), OECD Digital Education Outlook 2023: Towards an Effective Digital Education Ecosystem, OECD Publishing, Paris, https://doi.org/10.1787/c74f03de-en.
[16] OECD (2023), “Public policies for effective micro-credential learning”, OECD Education Policy Perspectives, No. 85, OECD Publishing, Paris, https://doi.org/10.1787/a41f148b-en.
[11] OECD (2019), Getting Skills Right: Making adult learning work in social partnership, http://www.oecd.org/employment/emp/adultlearning-work-in-social-partnership-2019.pdf.
[8] Red Seal Program (2024), Occupational Standards, https://www.red-seal.ca/eng/resources/n.4.1.shtml.
[24] Vaitkute, L. (2022), National curricula - dilemmas and opportunities: Presentation of Lithuanian approach, https://www.cedefop.europa.eu/files/05._breakout_room_2_-_national_curricula_dilemmas_and_opportunities_lithuania_-_lina_vaitkute_1.pdf.