There is no shortage of articles telling academics that Generative AI is going to transform education. It is, and it will continue to do so. However, many of these pieces are long on enthusiasm and short on detail. This is not one of those.
What follows is a practical account of using ChatGPT to build a real teaching session. I’ll cover what I did, what worked, what failed, and how long it actually took. No hype—just the reality of how it saved me time and how it could possibly do the same for you.
The Test Case
My subject was a four-hour session on Pytest in Django, aimed at final-year BSc Software Engineering students. These students have a basic grasp of Django but possess solid overall coding skills. The session was split into a one-hour lecture and three hours of hands-on practical work in VS Code.
The Strategy: Starting with the Prompt
The key to getting useful output is being specific upfront. Rather than simply asking ChatGPT to "create a lesson on Pytest," I provided a detailed prompt specifying the audience, topic, format, and—crucially—how I wanted the interaction to work. I wanted an iterative process where the AI asked me questions until I was satisfied before producing the final content. Here is the prompt I used:
"I want to create a four-hour teaching session — one hour lecture and three hours of practicals. Topic is Pytest in Django for a group of final-year BSc Software Engineering students. They have a basic understanding of Django.
I want this to be done in two parts: the lecture slides and then the practical teaching material using VS Code.
We will start with the slides. Please ask me questions until I type 'now stop,' at which point the slides should be produced. Then we will move to the practical material, again iterating through questions until I type 'done now'."
This two-part structure was deliberate. By separating the lecture from the lab, I ensured each section got the focus it deserved without the AI trying to do everything at once.
The "Thinking Partner" Approach
What worked best was the question-and-answer refinement loop. Instead of generating a wall of generic content, ChatGPT asked clarifying questions about learning objectives, the depth of detail required, and the specific tools students would use.
This is the part most guides skip: GenAI tools are far more effective as a thinking partner in the design phase than as a one-shot content generator. The questioning actually helped me think through what I did—and didn't—want to include, which ultimately helped me do a better job.
The Results: What was Produced?
The Lecture Slides: The initial output provided a logical structure: testing concepts, Pytest vs. Django’s built-in runner, fixtures, and mocking. However, it struggled to calibrate the depth. The first pass was pitched at beginners; it took a few rounds of the question loop to bring the content up to the level of final-year students.
The Practicals: ChatGPT produced a series of stepped exercises. The structure was a useful scaffold, but the exercises initially lacked context. They were "bare basics" and needed more "why" behind the "what" to be truly educational for these students.
The Reality Check: What I Changed
The code examples required the most intervention. While some were fine, others contained small but meaningful errors. These are the "silent killers" of a teaching session—errors that would waste ten or twenty minutes of lab time while students struggle to figure out why their environment isn't running.
The Rule: Treat every AI code example as untested until you have run it yourself. I rewrote several examples substantially and tweaked others.
I also found the slide text a bit "flat." It was accurate but dry. I rewrote the explanatory paragraphs in my own voice to ensure the materials felt like they came from a human, not a manual.
The Bottom Line: How long did it take?
Building a session like this from scratch—slides, practicals, code examples, and timing—usually takes me six to eight hours.
Using this AI-assisted approach, the entire process took about 4 hours. That included the iterative questioning, reviewing the output, fixing the code, testing the code, and rewriting the text.
The time spent was cut by roughly 50%. However, that remaining time required your attention, and having a 'partner' asking meaningful questions helped as the activity changed. The saving is real; the shortcut is not.
Is it worth it?
| Good For... | Not Good For... |
| Structure: Getting a solid framework quickly. | Context: Understanding your specific students. |
| Ideation: Prompting you to think of missed topics. | Subject Knowledge: It cannot replace your expertise. |
| Mechanical Tasks: Saving time on slide building. | Accuracy: Producing ready-to-use code. |
| Scaffolding: Generating a base for exercises. | Calibration: Getting the pacing right without your input. |
Where to go from here
If you want to try this, start simple. Pick one session you are already planning. Write a prompt that specifies:
Your audience and their level.
The format you need.
The interaction style (ask me questions first, output second).
Review everything with the same critical eye you’d apply to a textbook you’ve never used before. Fix what’s wrong, cut what doesn’t fit, and keep the AI asking questions until you’re happy.
The goal isn’t to hand your job over to an AI. It’s to spend less time on the mechanical parts of the job so you have more time for the parts that actually require your expertise. In my experience, that is a trade well worth making.
Comments
Post a Comment