For the past few weeks, I’ve been exploring ways that educators can take back control of generative artificial intelligence. As part of curriculum design, I started with reflections on how I use multimodal generative AI myself when designing my blogs on education and resources.
Then, I took a brief side step into EdTech application design to experiment with ways educators can build their own AI assistants. Last week, I finished up with a critique of both edtech lesson planning and education think tanks who claim what teachers need is pre-made resources.
In my experience, educators are using these technologies in much more interesting ways than the options provided by the likes of Magic School and Khan Academy. With millions of dollars being invested nationally in artificial intelligence and in teacher workload, I want to shift the attention away from lesson planning as an administrative burden and towards curriculum design as a fundamental part of the profession.
In this post, I’ll outline a framework for using generative artificial intelligence in a way which values and relies on teacher expertise.
State and federal ministers for Education, if you’re reading this, this is how the hundreds of teachers I’ve spoken to in the last couple of years imagine artificial intelligence contributing to lesson planning and resource creation. As far as workload is concerned, we can save those button clicks for the actually burdensome administrative tasks that I wrote about in the previous article.
Developing the Framework: Design, Refine, Create
Last year, Mike Perkins, Jasper Roe, Jason MacVaugh and I published the AI Assessment Scale, a five-level scale to promote transparency in student use of artificial intelligence in assessments. A number of people have also observed the scale can be applied to how educators are using AI. For this curriculum design framework, I began by focusing on levels two, three and four: idea generation, editing and AI plus human evaluation, respectively.
You can access a free ebook on the AIAS with over 50 activities for the 5 levels by signing up for the mailing list here:
I also drew on other commonly used curriculum, design and teaching and learning frameworks, including Understanding by Design, Beverly Derewianka’s teaching and learning cycle (which informed our own writing cycle in Practical Writing Strategies), and design thinking instructional design frameworks like ADDIE.
I also applied some of my earlier work on the importance of design, as opposed to delivery in curriculum contexts.
Finally, and most importantly, I looked at how the educators I’ve been working with for the last two years are using generative AI in their day-to-day work. This includes the six original areas from Practical AI Strategies:
- Planning
- Refreshing
- Improvising
- Personalising
- Collaborating
- Communicating
And my more recent updates:
Throughout this work, I have been interested in how educators are using powerful foundation models such as GPT, Claude, and Gemini, and the major commercial applications built on top of them, like Microsoft Copilot. I’m also interested in the kinds of prompts educators use, the tasks they complete with AI, the challenges they face, and the resources they draw on.
In doing so, I’ve ended up with a three-part framework: Design, Refine, and Create. Here’s how each aspect of the DRC framework breaks down.
Design
As far as I’m concerned, curriculum design is where it all begins. As I’ve written elsewhere, design isn’t merely planning – it’s the deliberate crafting of an approach using all of the available resources, including expertise, technology, existing materials, and leaning on our colleagues. So it’s fitting that the first part of the DRC framework focuses on this fundamental aspect of teacher work.
Educators are using generative AI as part of the design process in many ways, including:
- Generating ideas for units of work
- Making connections between existing units and curriculum updates
- Developing cross-curricular connections
- Drawing on multiple resources and scope and sequence documents
- Transcribing faculty meetings and capturing curriculum planning discussions
- Bouncing ideas off a virtual colleague
- Role-playing different perspectives
- Synthesising multiple curriculum documents and resources
The key part of the design stage is that educators don’t want artificial intelligence to do all, or even most, of the work for them. A popular narrative for tech companies is that AI can remove the “fear of the blank page” and do the first draft – something I’ve questioned in an earlier article.
When I talk to educators, they generally find these AI first drafts supremely unsatisfying. The argument that it’s just a starting point, and then you work with it, falls flat in terms of reducing teacher workload when you consider the amount of time and effort required to get something useful from an AI platform.
But there are occasions when teachers might need artificial intelligence to generate some of the initial ideas, particularly in situations like early career teaching, where a lack of quality mentoring means the transition from university into the classroom can be incredibly difficult, or out-of-field teaching, which is rife in Australia and other jurisdictions, and means that teachers are often required to teach unfamiliar materials with limited resources.
Even in these cases, teachers are finding artificial intelligence doesn’t provide the answers. In my own PhD studies, one participant, an early career teacher, described trying to use AI to create essay topics, finding them “weirdly unsatisfying” and being wary of using them, because, by her own admission, she didn’t yet have skills to identify exactly why the essay topics didn’t work.
So the design stage is not about offloading idea generation or planning onto AI. It’s about pulling resources together to create new or updated materials built on significant prior knowledge and expertise.
Refine
The second aspect of the framework is Refine, where educators are using generative AI to edit, revise, seek feedback on and update teaching and learning materials. This aspect includes differentiation and modification, taking existing materials and using generative artificial intelligence to create multiple versions of a single resource.
In my experience, there are many ways educators are using generative AI to refine their materials, including:
- Combining old and new resources
- Synthesising ideas from multiple colleagues (for example, three or four members of a faculty working on a complete unit)
- Editing communications such as emails and newsletters and website copy
- Reviewing curriculum materials for diversity and inclusion
- Swapping out units of work (for example, by taking an existing unit on a given text and changing the core text being studied)
- Reviewing lessons for feedback and adjustment
- Transcribing professional learning and using the transcription to update Teaching and Learning Materials
- Reviewing academic literature to incorporate research
Like the design aspect, when educators use AI to revise and refine, it is less about generating content with artificial intelligence and more about using the AI as an interlocutor, an intermediary between the educator’s first draft and the final output.
Create
When unconstrained with the kinds of administrative burdens we would actually like to offload onto artificial intelligence, educators are incredibly creative. This aspect of the framework looks at ways in which educators are not only creating documentation, but are also using the technology in exciting, innovative ways.
In the context of workload, allowing the space and time for this kind of experimentation could contribute to teacher self-efficacy and feelings of confidence with the technology. Solutions which have simplified AI by turning it into a web page full of single-use buttons deny educators the opportunity to be really creative. But there are many teachers pushing back against the simplifications of the technology.
In the context of curriculum design, this includes:
- Creating meaningful resources with generative artificial intelligence
- Using image, audio and video generation tools to create multimodal materials
- Building custom chatbots and projects to assist with the design and creation of the implements and resources
- Creating customised tools for student use
- Creating novel ways to teach with and through artificial intelligence
Creating with generative AI tools is very different from “generating”. “Generate” suggests a lack of agency, a mechanised process where a teacher simply hands off the responsibility of contemplation.
I’m sure some of will have read articles exploring whether AI output can be creative, or even arguing that it outperforms humans in certain aspects of creativity, but the general consensus is that AI is not, in and of itself, creative, nor does it democratise creativity. What it does, in the language of multimodal design I was using earlier, is provide additional resources and affordances which teachers can add to their existing methods.
As these technologies become integrated into the software we use daily, such as Microsoft 365, Adobe products, Google Workspace, educators will find even more ways to create meaningful resources, but only if they are given the time and professional learning required to do so.
The Final Framework
Having gathered up examples of how educators are using generative AI, there’s one final important aspect to the DRC framework which draws it closer to design thinking frameworks used elsewhere in education. And that is to add an element of iteration. Because in curriculum design, the idea of generating a single lesson plan or even a single unit of work is fairly pointless. Individual lessons, units, or even entire curricula cannot be decontextualised from the whole.
Clicking a button to create a lesson plan suggests that lessons are like individual Lego bricks scattered in a hap-hazard pile across the ground. Curriculum design is more like a complex and elaborate Lego set where each part has a specific role to play. The instruction manual might be difficult to follow at times, but the end product is always satisfying.
So in the final framework, it’s necessary to add some feedback loops and iteration. For example, a teacher may use generative AI to design a piece of work and then pass it along to other colleagues for review before revisiting those initial ideas. Or they could go from design to the creation of resources, and then find out that at the coalface, the resources fell flat. Back to the design drawing board.
Teams can cycle through the process of refining and creating materials, adapting them for different contexts, different learners, perhaps reusing materials across year levels or across subjects in different contexts.
Importantly the DRC framework is not a curriculum design framework in and of itself: as I’ve hopefully made clear throughout this article, GenAI cannot be used to create meaningful resources without guidance and expertise. The framework should be used alongside your existing curriculum design methods, whether that’s backwards design, ADDIE, or something else.
So with all that in mind, here is the final framework:
Over the next few weeks, I’ll continue to provide examples of how educators are using generative AI in these aspects of the framework.
Please leave comments here or on LinkedIn. Engage in the discussion, critique, offer suggestions and help to move us beyond simplistic uses of this incredibly powerful technology.
Want to learn more about GenAI professional development and advisory services, or just have questions or comments? Get in touch:
Your message has been sent
Leave a Reply