Prompting Progress: How Artificial Intelligence is reshaping the classroom
This example is shared by a Modern Studies teacher from Graeme High School in Falkirk Council and uses Gemini and NotebookLM
Use case: Transforming Differentiation in a Modern Studies Classroom
I’ve been using Google tools to make my Modern Studies lessons more accessible for the range of learners I’ve got in front of me. NotebookLM has been particularly useful for bridging the gap between complex topics and getting students actually engaged with them. I create short overview videos that introduce new units or work as revision prompts. I share these in our class Teams and use them as lesson starters. They give students a visual and auditory “brain nudge” before we get into deeper source analysis or critical thinking. It builds confidence when they’re not going in cold. The infographic feature is brilliant for breaking down tricky Modern Studies concepts or specific SQA exam skills. It gives students a low-stakes way into topics that would otherwise be overwhelming walls of text.
Gemini has become essential for differentiation. I use it to adapt reading passages and web content into tiered worksheets aligned with SOLO taxonomy. This means students can work through questions at their own pace and level – proper support for those with Additional Support Needs, and meaningful challenge for those ready for extension work. Recently I used Gemini to adapt a class test for students with significant literacy barriers. I simplified the sources and question then generated specific word banks and used the image creation feature to provide visual scaffolds. These helped students recall key information and actually articulate what they knew. The point is assessing their understanding of the political or social issues, not just their reading speed or vocabulary.
Crafting prompts
How I prompt the AI depends on what I’m creating, but I always tell it to consider dyslexic and neurodivergent learners, and students with English as an Additional Language. I ask it to check its own outputs against these needs and suggest adjustments for me to review. This isn’t just about simplifying text. I use it to inform decisions about colour contrast and images as well, making sure the final resources actually work for all students.
Crafting prompts takes tweaking to get something actually usable. Early attempts didn’t fit my teaching. I’ve learned to be specific and not “simplify this” but “adapt this for S2 students reading at S1 level, with visual supports for key vocabulary”. The more detail I give about what I actually need, the better it works. I also build in a checking stage. After it generates something, I ask “what assumptions did you make?” or “is this accessible for dyslexic learners?”. That catches problems before they reach students. Sometimes it takes a few rounds. You need to generate, spot what’s wrong, adjust the prompt, try again. But even with going back and forth, it’s still faster than doing it all manually. The tweaking has actually improved my practice. Being forced to articulate exactly what reading level or scaffolding I need has improved my thinking about differentiation generally. Over time I’ve figured out prompts that work reliably for tasks I do regularly, so those are quicker now. New stuff still needs experimenting.
After it generates something, I ask “What assumptions did you make?” or “Is this accessible for dyslexic learners?”
How the tools improved the task I was doing
These tools have transformed differentiation from something I was doing but found time-consuming, into something that takes minutes instead of hours. I was already creating resources for different learners, this just makes it quicker and more sustainable. I can create tiered resources, simplified sources, and visual scaffolds quickly enough to do it for every lesson, not just highstakes assessments. This means students with literacy barriers, EAL students, and neurodivergent learners can access the same concepts as their peers without losing challenge for more confident students. The time saved lets me focus on the pedagogy rather than mechanically rewriting text. I’m making better decisions about scaffolding and assessment because I’m not exhausted from creating multiple versions manually. The real impact: differentiation is now sustainable and consistent, not occasional.
Next steps
My next steps are to get students thinking about they use their own AI to help themselves.
I know they are using it but I want to help them make the best use of it as well.
Examples
Click the NotebookLM examples below to enlarge.
N5 Assignment
Parents guide to AI toys
