UNESCO AI CFT: AI competences for teachers
A framework that does not start from tools
The UNESCO AI Competency Framework for Teachers, published in 2024, defines 15 competences across 5 dimensions. What sets it apart is not the count, but the order the dimensions are arranged in.
You would expect a framework on artificial intelligence for teachers to start from tools: which systems exist, how to use them, what they can do in class. UNESCO does the opposite. The first two dimensions concern a human-centred mindset and the ethics of AI. The tools come afterwards.
The five dimensions
- Human-centred mindset -- recognising that AI is a tool serving human choices, and that responsibility for educational decisions stays with the teacher.
- Ethics of AI -- privacy, bias, equity of access, transparency; the ability to recognise when a system produces discriminatory effects.
- AI foundations and applications -- how an AI system broadly works, what it can and cannot do, what its structural limits are.
- AI pedagogy -- integrating AI into instructional design rather than bolting it on; assessing learning in a context where students have access to generative systems.
- AI for professional development -- using AI for your own continuous professional growth.
Three progression levels
Each dimension is articulated across three levels: acquire, deepen, create. Hence the total of 15 competences -- five dimensions by three levels.
The progression runs from informed use of existing tools to the ability to rethink your own teaching practice, and on to actively shaping how AI enters the school. It is not a scale of technical skill: it is a scale of professional agency.
Why the order is the important part
A training pathway starting from tools produces teachers able to use a generative system and lacking criteria for deciding when not to. The criterion is not an accessory to technical skill: it is its precondition.
This is the same logic GreenComp applies to sustainability, putting values before action. That is not a coincidence: it is what separates a competence framework from a user manual.
The relationship with DigCompEdu
The two frameworks do not overlap, they compose. DigCompEdu describes a teacher's digital competences as a whole, across six areas and twenty-two competences. UNESCO AI CFT goes deeper on one specific dimension that, when DigCompEdu was written, did not have the centrality it has today.
In practice: DigCompEdu remains the grid for planning a school's training overall; AI CFT is the instrument for the artificial intelligence part of it. DigComp 3.0 has likewise integrated AI competences for citizens, with the same distributed approach.
Browse the UNESCO AI CFT framework and its relationships with the other European frameworks in the framework catalogue on ExplorerHub.
Related articles
FinComp and OECD/INFE: which reference to cite
They are not two rival frameworks: FinComp comes out of the collaboration between the European Commission and the OECD. But in a project it pays to know which one to cite, and why.
LifeComp: personal, social, learning to learn
Nine competences across three areas, 27 descriptors. LifeComp is the smallest framework in the European landscape, and that is exactly what makes it the easiest to use as a cross-cutting grid.
FinComp: the European financial competence framework for adults
Developed by the European Commission with the OECD, FinComp describes 22 competences across 4 areas. It does not teach investing: it describes what you need to know, to do, and how to be disposed towards money.