From AI Use to AI Competence
Why Education Systems Must Act Now and how AIComp can support

von Ulf-Daniel Ehlers  |  27. Januar 2026

OECD Digital Education Outlook 2026: Exploring Effective Uses of Generative AI in Education

The OECD Digital Education Outlook 2026 leaves little room for ambiguity: generative AI is already deeply embedded in educational practice, yet its impact on learning remains fragile, uneven, and highly dependent on pedagogical design (OECD, 2026). Students across vocational education and training (VET) and higher education use AI tools extensively—but too often as shortcuts for task completion rather than as instruments for competence development (OECD, 2026).

The central insight of the OECD report is both sobering and clarifying: successful task performance with AI does not automatically translate into learning. On the contrary, when AI replaces learners’ cognitive effort, metacognitive engagement declines, and durable competence development is undermined (OECD, 2026). This diagnosis challenges education systems at their core.

What follows from this is not a call for restriction or prohibition—but for conceptual clarity and curricular transformation, with a particular focus on integrating AI competence as a foundational aspect of VET and higher VET curricula.

---

Why We Need Clear Concepts of AI Competence

One of the OECD report’s strongest messages is that education systems urgently need clear and shared understandings of what AI-related competencies actually are (OECD, 2026). Without this clarity, institutions risk oscillating between uncritical adoption and defensive resistance.

Crucially, AI competence must not be reduced to technical proficiency or tool usage. The evidence presented by the OECD shows that how AI is used matters far more than whether it is used. Learning gains emerge when AI is embedded in pedagogical designs that foster cognitive activation rather than cognitive offloading, reflection instead of automation, and learner agency instead of dependency (OECD, 2026).

This aligns closely with the AIComp project’s competence model, an empirically grounded, structured framework of Artificial Intelligence Competences that identifies twelve competence fields across professional, personal, and social domains, reflecting the future skills individuals need in an AI-shaped world (Ehlers & AIComp Project Team, 2024).

AIComp serves not as a checklist of tool proficiencies, but as a reference model for action competence—capabilities that integrate knowledge, skills, ethical disposition, and self-regulated agency (Ehlers et al., 2024).

Fig.1: AICompe Model (www.ai-comp.org)

From Competence Models to Curriculum Transformation

Competence models alone, however, are insufficient. The OECD report repeatedly emphasises that general-purpose AI tools rarely align with curricular structures and that meaningful learning outcomes depend on intentional curricular and pedagogical design (OECD, 2026).

This leads to the second, equally pressing challenge: How can curricula in VET and higher education be transformed to integrate AI competencies systematically and sustainably?

One powerful answer lies in the agile curriculum development approach that AIComp supports. Rather than treating AI-related content as optional or supplementary, institutions should embed AI competence development across qualification profiles (Ehlers et al. 2024).

At the heart of this process is a three-step model for implementing competence frameworks into curricula:

---

Implementing Competence Frameworks in Curricula:

 

A Three-Step Approach

 

1. Interpret the Framework as a Reference Model

Before practical application, a competence framework — whether AIComp or a broader reference tool — must be deeply understood as a reference model. This involves engaging with its structure, underlying assumptions, and the vision of the learner it promotes. Educators must ask:

• What types of action competence does this model prioritise?
• How are responsibility, reflection, and agency embedded in the competence descriptions?

This analytical phase enables responsible curricular adaptation rather than superficial alignment (Ehlers & AIComp Project Team, 2024).

2. Contextualize the Framework for the Learning Environment

After interpretation, the framework must be contextualised to the specific educational setting:

• Selective Adaptation – Identify which AIComp competence fields are most relevant to a given VET or higher VET programme.
• Thematic Specification – Translate general competence statements into concrete learning objectives fitting the disciplinary and professional context.

For example, “ethical reasoning in AI use” might become “evaluate ethical implications of automated decision-making in healthcare diagnostics” in a nursing VET curriculum (Ehlerset al., 2024).

3. Design Learning Experiences for Competence Acquisition

Finally, educators design learning experiences that enable students to develop these competences. This requires:

• Choosing appropriate social forms (e.g., team collaboration, peer feedback);
• Selecting learning formats (e.g., project-based challenges, simulations, flipped classrooms);
• Formulating tasks and problems that require active decision-making, reflection, and responsible AI use.

This step transforms abstract competence descriptions into embodied, situated learning processes (OECD, 2026).

---

Two Concrete Examples of AI Competence Integration (Based on AIComp)

 

Example 1: Integrating “Critical Digital Competence” into a VET Mechatronics Programme

 

• Interpret: Using AIComp as a reference, the educational team identifies critical digital competence as essential for mechatronics professionals who will work with AI-enabled industrial systems (Ehlers et al., 2024).
• Contextualize: The team formulates specific learning outcomes such as “Analyze the outputs of adaptive manufacturing systems and identify bias or unintended automation errors.”
• Design Learning Experiences: Learners engage in project-based tasks where they compare AI-generated diagnostics with manual assessments, discuss limitations, and reflect on ethical ramifications of automation decisions.

This design goes beyond tool use to develop judgment, reflection, and agency in professional contexts (OECD, 2026).

Example 2: Embedding “Communication and Collaboration Competence” in Higher VET Business Programmes

 

• Interpret: Educators identify interpersonal and cross-disciplinary competences in AIComp as vital for future managers (Ehlers et al., 2024).
• Contextualize: They define outcomes like “Collaborate with multidisciplinary teams to design AI use cases that balance business goals and societal impact.”
• Design Learning Experiences: Students work on real partnerships with industry partners to co-design AI-assisted business solutions, documenting roles, decisions, and reflections on stakeholder needs and ethical considerations.

Here, competence development is grounded in real-world social interaction and collective decision processes (OECD, 2026).

---

From Curriculum Design to Study Programme Development

At the level of study programme development, AI competence integration must be structural, not marginal. AIComp supports this through playbooks, card sets, and praxis-oriented materials that help educators translate competence frameworks into programme architectures (Ehlers et al., 2024).

These materials guide institutions in iterative and collaborative design cycles, enabling them to test micro-pilots within existing structures and refine curricula based on learner feedback and competence development outcomes — precisely what the OECD highlights as crucial for effective AI adoption in education (OECD, 2026).

---

Conclusion

The OECD Digital Education Outlook 2026 makes one message unmistakably clear: the future of education in the age of AI will not be decided by technology, but by the competence frameworks, curricula, and learning cultures we deliberately design (OECD, 2026).

For VET and higher VET institutions, the task is urgent and active: move from AI use to AI competence by anchoring AI integration in well-defined, action-oriented competence frameworks like AIComp — and by translating these frameworks into learning experiences through a deliberate, three-stage implementation model (Ehlers et al., 2024).

---

References (APA 7)

Ehlers, U.-D., Lindner, E., Rauch, E. (2024). AIComp: A competence model and playbook for agile curriculum development of AI competencies. AIComp. https://www.ai-comp.org

OECD. (2026). OECD Digital Education Outlook 2026: Exploring effective uses of generative AI in education. OECD Publishing. https://doi.org/10.1787/062a7394-en