What 'AI Competent' Actually Means for a Teacher in 2026

I have lost count of the number of times a school leader has told me their teachers are "AI competent" because they have attended a training session. Usually it was a one-hour twilight on how to use ChatGPT. Sometimes it was a vendor demonstration for a new platform. Occasionally it was a conference keynote where someone showed impressive AI demos on stage.

None of these make a teacher AI competent. They make a teacher AI aware, which is a different thing entirely; and the gap between the two is where most schools' AI integration efforts fall apart.

The Problem with Undefined Competency

When a school says its teachers are "safeguarding trained," there is a shared understanding of what that means. There are levels, there are accredited courses, there is a progression from basic awareness through to designated safeguarding lead. Everyone knows what the standard is and how to measure against it.

No such shared understanding exists for AI competency in most schools. "AI trained" means whatever the school wants it to mean. In practice, it usually means "has sat in a room while someone talked about AI." There is no standard, no progression model, no competency framework, and no way of knowing whether the training actually changed anything about how teachers think or work.

This is a problem because teacher AI competency is the single most important factor in whether AI integration benefits students. The research is clear on this, and every major international framework, from UNESCO to the OECD, identifies teacher readiness as foundational. Not supportive. Not helpful. Foundational. As in: without it, nothing else works.

What the Frameworks Say

UNESCO's AI Competency Framework for Teachers is the most detailed attempt to define what teacher AI competency actually looks like. It maps competencies across several areas, including understanding AI fundamentals, applying AI in teaching and learning, evaluating AI critically, and engaging with the ethical and social dimensions of AI in education.

What strikes me about the UNESCO framework is how much of it has nothing to do with using AI tools. A significant portion of the competency model is about understanding: what AI is, how it works at a conceptual level, what its limitations are, and what its implications are for education and society. This matters because a teacher who can use ChatGPT but does not understand that it generates probabilistic text rather than factual information is not competent. They are fluent in a tool they do not understand, which is arguably more dangerous than not using it at all.

The OECD takes a complementary approach, focusing on the skills teachers need to navigate an AI-augmented professional environment: adaptability, critical evaluation, ethical reasoning, and the ability to redesign learning experiences that account for AI's capabilities and limitations.

Both frameworks make the same fundamental point: AI competency is not tool proficiency. It is a combination of knowledge, skills, and dispositions that enables a teacher to make informed, critical, and pedagogically sound decisions about AI in their practice.

A Practical Competency Model

Based on my work with schools and my reading of the international framework landscape, I use a five-level competency model that I find practical for school leaders to work with:

Level 1: Awareness. The teacher knows that AI exists in education, can name common AI tools, and has a basic understanding of what generative AI does. They have not meaningfully integrated AI into their practice.

This is where most teachers currently sit after a single training session. It is a starting point, not an endpoint.

Level 2: Exploration. The teacher has experimented with AI tools in their professional practice: using AI to generate resources, draft communications, or plan lessons. They can identify situations where AI is helpful and situations where it is not. They are beginning to develop a sense of AI's limitations.

This is where teachers typically arrive after sustained, supported experimentation; not after a single session, but after several weeks of guided practice with reflection.

Level 3: Integration. The teacher deliberately and thoughtfully integrates AI into their teaching and assessment practices. They can design learning activities that leverage AI appropriately, they can explain to students when and how AI should and should not be used, and they can adapt their assessment approaches to account for AI. They understand data protection implications and follow the school's AI policy consistently.

This is the target level for most classroom teachers. A school where the majority of teaching staff are at Level 3 has a genuinely AI-competent workforce.

Level 4: Evaluation. The teacher can critically evaluate AI tools: not just in terms of usability, but in terms of pedagogical value, ethical implications, bias, data practices, and alignment with educational objectives. They contribute to the school's AI governance by reviewing tools, advising colleagues, and informing policy development.

This is the target level for heads of department, CPD leads, and members of the AI steering group.

Level 5: Leadership. The teacher can design and deliver AI CPD for colleagues, contribute to whole-school AI strategy, and engage with the wider professional community on AI in education. They understand the international framework landscape and can position the school's approach within it.

This is the target level for your AI lead, your head of teaching and learning, and potentially your senior leadership team.

How to Measure It

A competency model is only useful if you can assess against it. I recommend a combination of self-assessment and evidence-based evaluation.

Self-assessment is the starting point. Give teachers the five-level model and ask them to place themselves. Self-assessment has known limitations, people tend to overestimate their competency, but it is valuable as a starting point and as a tool for building self-awareness.

Evidence-based evaluation provides the rigour. At each level, define what evidence would demonstrate competency. At Level 2, it might be a reflective log of AI experimentation. At Level 3, it might be a lesson plan that meaningfully integrates AI with a rationale for why and how. At Level 4, it might be a written evaluation of an AI tool against pedagogical and ethical criteria.

Run the assessment at the start of the academic year to establish your baseline. Design your CPD programme to target the most common gaps. Re-assess at the end of the year to measure progress. This gives you the data you need to demonstrate impact to your leadership team, your governing body, and your accreditation body.

The AI Literacy Audit Tool includes a teacher competency dimension that assesses your school's CPD provision against international framework expectations. It tells you not just where your teachers are, but where the frameworks say they should be; and highlights the gap.

What Good CPD Looks Like

If your current CPD programme is a series of one-off workshops, it will not move teachers along this progression. Research on professional development, not just in AI, but across all areas, consistently shows that effective CPD is sustained, practice-based, collaborative, and reflective.

Sustained means it happens over time, not in a single session. A term-long programme with regular touchpoints is far more effective than a standalone training day.

Practice-based means teachers are trying things in their classrooms, not just hearing about them in a training room. Every CPD session should end with a practical task that teachers implement before the next session.

Collaborative means teachers learn from each other. Peer observation, co-planning with AI tools, and department-level discussions about AI integration are all more powerful than isolated individual exploration.

Reflective means teachers are thinking critically about what works and what does not. Structured reflection, not just "how did it go?" but "what did the AI add to student learning that would not have been there otherwise?", drives deeper engagement.

The Stakes

I want to be clear about why this matters. The schools that get teacher AI competency right will be the ones that realise AI's potential to improve teaching and learning. Their teachers will use AI thoughtfully, their students will develop genuine AI literacy, and their governance structures will work because they are staffed by people who understand what they are governing.

The schools that treat competency as a box-ticking exercise, attendance records for training sessions, certificates for completing online modules, will have teachers who are AI-exposed but not AI-competent. The difference will show up in classrooms, in assessment integrity, in parent confidence, and in inspection outcomes.

The international frameworks are pointing in the same direction on this. Teacher AI competency is not optional. It is foundational. And defining what it actually means, clearly, measurably, and honestly, is the first step to building it.