As a History teacher, one strategy I use to help students think deeply is comparison. Sometimes this takes the form of a Venn diagram. For more senior students, however, I deliberately make the comparison less obvious. A Year 12 class might be asked to compare the political structure of Imperial Russia with a game of cricket.

By considering similarities and differences between seemingly unrelated ideas, students are forced beyond first recall. They begin to identify underlying concepts, question assumptions and view familiar knowledge from new angles.

Recently, I found myself applying this same approach to AI and education.

As I set out to become a more effective user of generative AI, I noticed the process felt remarkably familiar.

The more conversations I had with AI systems, the more I found myself drawing on habits I had developed over years in the classroom.

Good teaching is not simply presenting information. As teachers, we determine what a student understands, what information they need, how much support to provide, and what sequence might best help them progress. We frame questions and reword them when understanding fails. We probe. We challenge assumptions. We provide context. We diagnose misconceptions. We scaffold when support is needed and gradually withdraw it when it is not. We differentiate language, examples and level of challenge constantly.

All of this is iterative, happening in every moment of a lesson and in the hours of planning and reflection outside of class. Teaching is a continuous feedback loop of acting, interpreting, adjusting and trying again.

Fairly quickly, I realised this was largely how I was instinctively using AI. I rarely get what I want from one perfectly engineered prompt. I need to be clear about what I am trying to produce first. I need to provide context, set constraints, specify the audience, purpose, level and format. Sometimes I give examples. Often, I break the task into stages.

When the response misses the mark, I diagnose why. Has the model misunderstood the task? Have my instructions been ambiguous? Has it made an assumption I did not intend? Has it drifted into distraction? Am I expecting it to read my mind?

When outputs are produced, I ask interrogating questions: Why has this conclusion been reached? What evidence supports it? What perspective is missing? What would challenge this interpretation? Where is the reasoning weakest?

This process is one of constant formative and summative evaluation – something teachers constantly do in our classrooms every day.

Others have made similar observations. Jeremy Utley, an AI innovation expert and Stanford adjunct professor, has argued that teachers and coaches often work effectively with AI because they are skilled at asking questions, providing context, challenging assumptions, and guiding an interaction.

Reflecting on the comparison, the overlap in the Venn diagram between effective teaching and effective AI use is larger than it first appears. Teachers, in many respects, possess habits of thinking that transfer naturally to productive AI use.

This is why teachers are approaching AI with both interest and discernment. We recognise its potential, but we also understand that meaningful results depend on judgement, context, feedback, and continual refinement. We also know that while AI can be a powerful tool, the goals of education are not to simply produce more efficient outputs.

As we continue to explore its possibilities, teacher expertise remains one of our greatest assets.

Teachers bring professional judgement, disciplinary knowledge, and a deep understanding of learning to the decisions they make.

This article introduces a series exploring good pedagogy, specifically pedagogy in an AI-rich world. Through a range of classroom case studies, teachers will share how they draw on their professional knowledge and experience to make decisions in response to enduring pedagogical questions about what students should know, be able to do, and become. These questions sit at the heart of education and remain as important as ever.

Click here to continue to Part 1 of this series.

Claire Butler

Claire is a Research Fellow (AI in Education) with the Barker Institute, with a focus on developing research-informed, classroom-ready approaches to AI that strengthen thinking, assessment integrity, and teacher capability. She holds a Master of Education (Educational Studies) with Excellence from UNSW, with research centred on effective teacher professional learning. Claire has extensive experience in the classroom and in leadership roles across Sydney independent schools, including Pymble Ladies’ College, Abbotsleigh, and most recently as Head of History at The Scots College. A Modern History and History Extension teacher and HSC Senior Marker, Claire is a regular contributor to professional journals and has presented at state and national education conferences. Her current work sits at the intersection of AI literacy, assessment design, and instructional routines that help students use AI ethically and effectively.