Why Teachers Need More Than Technical Skills to Integrate AI in Education

By /

There’s a common assumption in conversations about AI in education: if teachers learn how to use the tools, they’ll figure out how to teach with them. The logic seems straightforward. Show them how ChatGPT works, walk them through a few prompts, and they’ll integrate AI into their classrooms.

Celik (2023) tested that assumption empirically and found it doesn’t hold up.

His study introduces the Intelligent-TPACK framework, an extension of the well-known TPACK model that accounts for the unique demands of teaching with AI. The findings show that technical knowledge alone does not predict effective AI integration in education. What predicts it is a combination of pedagogical judgment, ethical assessment, and subject-specific understanding of how AI tools connect to instructional goals.

Knowing How AI Works Is Not Enough to Integrate AI in Education

The structural equation model in Celik’s study reveals a clear pattern. Teachers who understand how AI tools function at a technical level are better positioned to assess AI decisions and recognize pedagogical possibilities. But that technical knowledge, on its own, does not lead to meaningful classroom integration.

As Celik puts it: “TK allows teachers to better assess decisions of AI. However, only TK is not sufficient educational integration of AI-based tools” (p. 8).

This finding challenges a lot of the professional development currently being offered to teachers. Many AI training programs focus heavily on tool demonstrations, prompt engineering workshops, and platform walkthroughs. These have value, but they address only one layer of what teachers actually need. A teacher can know exactly how ChatGPT generates responses and still have no clear idea of when, why, or whether to use it with their students.

Integrate AI in Education

I covered a related finding in a previous post on AI literacy for teachers, where Bilbao-Eraña and Arroyo-Sagasta (2025) showed that an 8-hour AI literacy intervention improved teachers’ awareness and attitudes but failed to build trust.

Trust required deeper engagement with ethics and governance. Celik’s data tells a similar story from a different angle: technical training moves the needle on awareness, but pedagogical and ethical knowledge are what actually drive integration.

Ethics Is Central to AI Integration in Education

One of the strongest contributions of the Intelligent-TPACK framework is how it positions ethics. In the traditional TPACK model, ethical considerations are either absent or treated as background context. Celik brings them into the core of the framework and tests their influence empirically.

The results are striking. Ethical assessment, defined as the ability to evaluate AI decisions for transparency, fairness, accountability, and inclusiveness, functions as a direct contributor to teachers’ pedagogical use of AI. Celik reports that “the ethical assessment is not regarded as a main component of TPACK; however, it is as effective as TPK in influencing TPACK” (p. 8).

That’s a significant finding. It means a teacher’s ability to evaluate whether an AI recommendation is fair, whether an automated grading decision is transparent, or whether an adaptive learning pathway treats all students equitably has as much influence on effective AI integration as their pedagogical knowledge does.

Celik also notes something I find particularly promising about the relationship between AI use and ethical awareness: “The interaction with AI systems could lead to an increase in teachers’ ethical assessment skills” (p. 8). In other words, working with AI can itself become a site for developing professional ethical judgment, provided the right conditions and support are in place.

Pedagogical Knowledge Remains the Strongest Predictor

Across all the pathways in Celik’s model, pedagogical knowledge emerges as the most powerful predictor of full AI integration. Teachers who understand how AI changes feedback, monitoring, personalization, and assessment practices are the ones most likely to use AI tools in purposeful, educationally meaningful ways.

This connects to what Guo et al. (2025) found in their year-long classroom study: it was the pedagogy, not the tool, that determined whether ChatGPT added value to student learning. When instructors redesigned their approach to emphasize collaboration and active exploration, student perceptions improved dramatically. The tool stayed the same. The teaching changed.

Celik’s framework offers a theoretical explanation for why that pattern keeps showing up across the research. Pedagogical interpretation of AI outputs, understanding what a recommendation means for a student’s learning trajectory, deciding whether to act on an AI-generated alert or override it, knowing when AI feedback supports a learning goal and when it undermines one, is the knowledge that makes AI integration purposeful.

The Teacher as Orchestrator

Celik’s framework redefines the teacher’s role in AI-enhanced classrooms. The metaphor he uses is the orchestrator: someone who interprets AI outputs, evaluates their consequences, and aligns them with pedagogical intent. This is especially important when AI systems produce opaque or automated decisions that directly affect students, such as performance predictions, grouping recommendations, or adaptive content sequencing.

As Celik concludes: “the orchestrator role requires a teacher’s not only technical but also pedagogical and ethical knowledge and skills” (p. 9).

That framing resonates with me. It moves the conversation away from “can you use the tool?” toward “can you make professional judgments about what the tool is doing and whether it serves your students?” Those are very different questions, and they require very different kinds of preparation.

What This Means for Teacher Training

If we take Celik’s findings seriously, teacher preparation for AI needs to look fundamentally different from what most institutions currently offer. Technical training is a starting point, but the curriculum needs to extend into pedagogical reasoning about AI affordances, ethical evaluation of AI decisions, and subject-specific applications that anchor AI use in disciplinary teaching.

Professional development programs that stop at “here’s how the tool works” are leaving teachers underprepared for the judgment calls they’ll face in AI-enhanced classrooms. The Intelligent-TPACK framework gives us a structure for building something more comprehensive, and the empirical evidence behind it makes a strong case for why we should.

Reference

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top