Skills and competencies have always been central to vocational education and training.  But VET has always been a rather backwater sector in education compared to the prominence of the universities and the overwhelming size of the school sector. Until recently that is. The hype around AI has turned attention to what new skills might be required in rapidly changing economy. And even more so with productivity being seen as critical for economic growth and for innovation.  It is into this noisy landscape that the new OECD Skills Outlook 2025 report, “Building the Skills of the 21st Century for All,” arrives, offering not easy answers, but a crucial, data-rich foundation for a more substantive conversation. The report is a dense, 264-page document, and I cannot claim to have read every page! But its findings have implications for practice in VET, particularly concerning the skills demanded by AI and the role of VET in fostering them. It raises deeper questions about the purpose of learning and the role of VET when routine tasks are increasingly automated. I don’t agree with all of the findings but it raises issues for the future of VET, both from an eco0nomic and educational viewpoint.

Motivation in the Age of Automation

The report’s most compelling argument is not about technology itself, but about human motivation. As AI and automation become more adept at handling routine information-processing, the report argues that “human value creation increasingly hinges on the desire to engage creatively with ambiguous, open-ended challenges.” In other words, the future of work belongs not to those who can simply follow instructions, but to those who possess the intrinsic motivation to go beyond the brief, to experiment, to innovate, and to solve problems that don’t have a clear answer.

The VET sector has never been just about training for task completion, rather adopting a  wider focus on understanding or mastery that allows learners to adapt and thrive. The OECD’s findings suggest that as AI automates the “what,” the focus should shift decisively to the “why” and the “what if.” The report cautions that a decline in engagement with complex tasks, even in low-stakes environments, could erode the competitiveness and societal advancement.

The report presents sobering evidence that performance in literacy and numeracy among adults has declined between 2012 and 2023. This isn’t just about skills erosion; it may signal a broader disengagement from challenging cognitive work. In a world where AI can handle the routine, this trend can be seen as deeply troubling. OECD argues that the willingness to work beyond standard requirements is becoming “the decisive factor in individual productivity and innovation.” For VET, this suggests wider role in cultivating curiosity, persistence, and the confidence to tackle the unfamiliar.

AI: A Bridge for Skills Gaps, or a more unequality?

Interestingly, the report suggests that generative AI could be a powerful force for equity. By automating some routine intellectual tasks, it has the potential to narrow the gap in basic information-processing skills, allowing learners from all backgrounds to engage with more advanced material. This presents an opportunity for VET to leverage AI as a tool that scaffolds learning; however the familiar mantra of freeing up educators to focus on mentoring, coaching, and facilitating higher-order thinking is becoming rather jaded.

However the report draws attention to the stark disparities that presently exist in our education and training systems. The single strongest driver of skills gaps is socio-economic background. The very adults who most need upskilling and reskilling are the least likely to participate in adult learning—a gap of over 40 percentage points between the highest and lowest educated. Only 19% of adults with below upper secondary education participate in non-formal learning, compared to 61% of those with tertiary education.

Even more concerning for VET is the type of training being accessed. The report finds that training often “reproduces the intergenerational transmission of disadvantage.” Learners from disadvantaged backgrounds are over-represented in courses focused on operating machinery or following protocols. Meanwhile, those from more privileged backgrounds are more likely to be in training for project management, organisational skills, and other areas that lead to greater career mobility. AI doesn’t operate in a vacuum; it operates within this existing system, and without intentional, systemic change, it risks amplifying these inequalities rather than mitigating them.

The barriers to participation are also telling. The report identifies that family obligations weigh more heavily on younger workers and women, while those in rural areas cite scarcity of suitable courses and inconvenient scheduling. These are not abstract policy problems; they are the daily realities that shape who gets to develop advanced skills for the AI age and who gets left behind.

What Skills Actually Matter?

The OECD framework identifies three core 21st-century skills: literacy, numeracy, and adaptive problem solving. But it’s the last of these that deserves particular attention in the context of AI. Adaptive problem solving is defined as “the capacity to achieve one’s goals in a dynamic situation in which a method for solution is not immediately available.” This isn’t about following a procedure; it’s about navigating ambiguity, drawing on multiple resources, and monitoring one’s own thinking—precisely the metacognitive skills that AI struggles to replicate.

The report also emphasises social and emotional skills: conscientiousness, emotional stability, open-mindedness, and the willingness to delay gratification. The ability to persist through difficulty, to invest in long-term professional development, and to commit to complex projects are seen by some as the attributes that distinguish human expertise from algorithmic execution.

VET can be seen as having a key role in developing adaptive problem solvers who can apply their domain expertise in unpredictable contexts. These can be illustrated as the plumber who can diagnose a novel fault, the care worker who can respond to a complex emotional situation, the technician who can troubleshoot an unfamiliar system—all of which. are skills beyond the routine applications of AI.

From Policy to Practice

OECD calls for “agile lifelong learning policies” that are embedded in broader economic and innovation strategies. The report underscores the need for accessible pathways that allow for mobility between vocational education, academic programmes, and adult learning. An emerging role for VET is in developing such complex learning journeys.

The Designer of Challenging Experiences: If motivation is the new premium, a primary role for VET professionals is to design learning environments that cultivate it. This means creating tasks that are resistant to simple automation, that demand critical thinking, collaboration, and hands-on problem-solving. It involves using AI not as a shortcut to the answer, but as a tool for inquiry – a partner for brainstorming, a simulator for complex scenarios, or a source of initial data that learners must then critique, validate, and apply. The OECD report highlights several countries already experimenting with AI-powered language learning tools, from Canada’s “Voilà Learning” platform to Iceland’s computer-assisted pronunciation training. These initiatives show promise, but they work because they are embedded within pedagogical frameworks that emphasize human interaction and critical engagement.

The Facilitator of Lifelong Mobility: Lifelong learning is not a new idea but boundaries between education and work are becoming more fluid. This means guiding learners through a system of micro-credentials, upskilling opportunities, and flexible pathways that the OECD identifies as critical for workforce adaptability. It requires a deep understanding of the labour market and a commitment to helping learners build a portfolio of skills that is both broad and deep. The report emphasizes that effective lifelong learning frameworks must integrate education at all stages, from early childhood through to ongoing adult training. For VET, this means thinking beyond the immediate qualification and considering how programmes fit into a learner’s entire career trajectory.

The Critical Interpreter of Technology: An AI can’t understand the nuances of a specific trade or the ethical dilemmas of a particular profession. A carpentry instructor is best placed to determine if an AI design tool is a useful aid or a dangerous crutch. A healthcare trainer must guide students on the ethical use of AI diagnostic tools, understanding their biases and limitations. VET teachers and trainers have subject-matter expertise which is necessary for critically evaluating and contextualizing AI tools for specific domains. This requires not just technological literacy, but pedagogical and professional expertise that can distinguish between tools that enhance learning and those that merely simulate it.

The Challenge Ahead

The OECD Skills Outlook 2025 report makes clear that the skills needed for the 21st century are not just technical competencies, but the capacity to learn, adapt, and engage with complexity throughout a lifetime. For VET this is both challenge and an opportunity. VET is not preparing learners for a single job or even a single career; but equipping learners  with the foundations to navigate an uncertain future with confidence, creativity, and critical judgment.

The AI age will not be shaped by the technology developers alone. It will be shaped by the educators in workshops, training centres, and classrooms across Europe who make daily decisions about what to teach, how to teach it, and why it matters. That is where the real work begins.

About the Image

At whose expense do tech development and tech companies thrive? The hands in the image suggest those whose lives and livelihoods are impacted by AI – including but not limited to ghost workers, students, those suffering from abusive deepfakes and AI-generated scams, residents of areas impacted by datacenters, those on the job market facing displacement by the promise of automation, and those seduced by the ELIZA effect of chatbots to make decisions that negatively impact their health and wellbeing. The water and wave imagery evokes the rising tide of AI hype (the “AI bubble”) as well as the environment