Using AI to evolve a Curriculum: What Does It Actually Look Like?

There are plenty of articles explaining how AI can save teachers time. Most focus on writing prompts, creating worksheets or generating lesson plans.

After spending the last few months redesigning parts of our curriculum with AI, I'm not convinced that's the biggest benefit.

What I think AI does exceptionally well is help teachers analyse, question and continually improve a curriculum, but that process looks very different from the one-prompt success stories that often appear online.

That's what this series is about.

Over the last few months I've been redesigning key parts of our Key Stage 3 Computing curriculum with AI. At the moment that's four or five units, but the long-term ambition is much bigger. Rather than reviewing a curriculum every few years, I want it to become something that continually evolves through classroom experience, feedback and AI-assisted review.

One thing became obvious very quickly. AI wasn't particularly good at redesigning a curriculum until it understood the one we already had.

Before asking it to improve anything, I spent a considerable amount of time building context. That meant uploading PowerPoints, schemes of learning, curriculum maps, exam specifications and supporting resources. Only once AI understood how the curriculum fitted together did the quality of its suggestions really improve.

That, for me, has been the biggest lesson. Context matters far more than prompts.

This also explains why people often experience hallucinations. If AI doesn't have enough information, it fills in the gaps. The solution isn't necessarily a better prompt. More often, it's providing better evidence.

The other thing that surprised me was just how iterative the whole process became. We didn't redesign a unit in a single conversation. Some units took thirty or more discussions before we were happy with them. Our new Year 8 Cyber Security unit has already gone through two significant iterations before it's even been taught.

I don't see that as a weakness.

In fact, I think it's one of the biggest opportunities AI offers. Traditionally, we design a curriculum, teach it for a year or two and then review it. AI makes it much easier to revisit units whenever our thinking changes, technology moves on or classroom feedback highlights something that could be improved. Curriculum development becomes continuous rather than cyclical.

One unexpected outcome of this process was something we've started calling learning moments.

Instead of designing lessons slide by slide, we began designing them as a sequence of short learning cycles. Each learning moment starts with an interesting question that gets students thinking. We then teach and model the new concept before students complete a short mini mission to apply what they've just learnt. That cycle is repeated throughout the lesson.

It wasn't something we planned from the beginning. It emerged naturally through repeatedly questioning lessons, discussing alternatives and refining the curriculum. It's now become the backbone of every lesson we've redesigned.

One final observation surprised me.

AI hasn't reduced the amount of thinking involved in curriculum design. If anything, it's increased it. Every suggestion has to be challenged, justified and refined. The difference is that those conversations happen over a few evenings rather than being spread across months of meetings and planning time.

This series will document that entire journey.

Rather than sharing isolated prompts, I'll show how we analysed the existing curriculum, built a new scheme of learning, developed lesson resources and then refined them through further iterations. As the Year 8 Cyber Security unit is completed and taught, I'll also share the finished schemes of learning, lesson resources and reflections so the series becomes a genuine worked case study rather than simply another discussion about AI.

The aim isn't for people to copy our curriculum.

It's to share a process that can be adapted to any subject.

If there's one message I hope comes across, it's this:

The value of AI isn't in producing a curriculum quickly. It's in making it easier to question, challenge and continually improve the curriculum you already have.

Planned Series

Blog 1 – Using AI to Redesign a Curriculum: What Does It Actually Look Like?

An overview of the project, why context matters more than prompts and how AI became another member of the curriculum design team.

Blog 2 – Building Context: Teaching AI Your Curriculum

How I used existing PowerPoints, schemes of learning, curriculum maps and specifications to build enough context for AI to make meaningful suggestions.

Blog 3 – Creating a Scheme of Learning with AI

Turning curriculum analysis into a coherent scheme of learning with progression, retrieval and assessment built in from the start.

Blog 4 – Designing Lessons Around Learning Moments

How repeated iteration led to a lesson structure based around learning moments: question, think, teach, model, mini mission and review.

Blog 5 – Creating Lesson Resources with AI

Using AI to build lesson slide decks, questioning, images, activities and worksheets while keeping the teacher firmly in control.

Blog 6 – Continuous Curriculum Improvement

Using AI alongside classroom experience and student feedback to refine lessons before and after they're taught, creating a curriculum that continually evolves rather than waiting years for a formal rewrite.