All notes & resources

Building AI practice with teachers

My approach to a school AI trial: begin with a curriculum need, welcome questions, and examine learning and the full workload together.

When a teacher asks whether a new tool will improve learning or add work, I want that question in the conversation early. During the Seesaw introduction at International School Manila, a colleague’s question helped me recognise that I had not explained the learning purpose clearly enough. I would now start with a concrete classroom need and invite those questions earlier.

I bring that lesson to planning around AI. My current university work includes staff professional learning and AI assessment and governance groups. The approach below describes how I would work with a school team exploring a new use of AI.

Start with the subject and the learners

I would begin with an existing unit and the teacher’s judgement about where pupils need support. What knowledge and reasoning are they working towards? What do their current explanations show? Where is language, access or preparation getting in the way?

My contribution would be to bring options, prepare examples and agree useful support. A teacher might want a resource to adapt, a demonstration, modelling or co-teaching. Subject colleagues should help shape the learning goal, task and assessment from the beginning.

Welcome the questions that make the design better

Concerns about authorship, independent thinking and workload deserve a practical response. We can agree what assistance is permitted, what pupils need to explain themselves and where there will be an opportunity to work without AI. Access support should remain in place when we check independent understanding.

I would ask colleagues to challenge an example before trying it: does the summary remove uncertainty, does the task still require subject knowledge, and is the checking manageable? Feedback should be able to change the plan. My Cognitive Coaching and Adaptive Schools training supports how I listen, facilitate discussion and invite reflection.

Try one sequence and examine the next piece of work

A possible starting point is a source-comparison activity. After teaching the background knowledge, a teacher could use a clearly labelled, checked AI account alongside authentic sources. Pupils would justify corrections, then later interpret a different source independently.

That is an illustration for team planning. The teacher would choose the sources, vocabulary, challenge and success criteria. We would compare pupils’ explanations and look again later to see what they can apply. One successful activity would be a reason to investigate further, rather than enough evidence for a claim about the whole school.

The same judgement informed my MYP Design teaching: prototype logs exposed a missing step, so I paused production and taught students how to use testing evidence. In the ISM plant inquiry, spoken explanations helped us distinguish a barrier to expression from a gap in understanding.

Count the checking as part of the work

If a proposed benefit is time saved, I would include preparation, checking, corrections, setup and support in the comparison with usual practice. Faster drafting is only part of that picture. A learning or access benefit might justify extra effort, but the team should be able to discuss that cost openly.

Before extending a trial, I would ask whether a colleague can adapt it independently, whether pupils could participate fairly and what the work actually shows. We should be able to simplify, revise or stop an approach, with a clear owner for any shared resource.

Build a curriculum beyond individual tools

AI literacy needs a progression that subject and phase teams can use. UNESCO’s student framework provides a reference for breadth, including human judgement, ethics, techniques and system design. It does not replace a school’s own sequence or assessment decisions.

The EEF’s implementation guidance also offers a useful way to think about the people, context and process around a change. I would use those references alongside local policy and pupil evidence, then let colleagues’ experience shape the next step.

For examples of my existing staff-learning work, see Helping colleagues make it their own and my Jisc presentation and resources.