Dr. Olgun Sadık

If you haven’t read my earlier posts, here’s a quick recap: As part of IU’s Information Technology Services Generative AI Project, I’m testing enterprise editions of ChatGPT, Microsoft Copilot, and Google Gemini to help design a brand-new Professionalization and Ethics course for engineering students. In the first task, I asked each tool to create course learning objectives and revise them for a class of 60 students. This follow-up task focused on content design.

The prompt I gave each tool was simple:

“Let’s focus on the first objective for this context and create an outline of the content to meet that objective.”

Here are the first learning objectives from each tool:

• ChatGPT: “Recognize and explain core principles of engineering ethics and professional conduct.”
• Microsoft Copilot: “Identify and explain core ethical principles and professional responsibilities relevant to engineering practice, including codes of ethics from major engineering societies (e.g., NSPE, IEEE).
• Google Gemini: “Students will be able to identify, explain, and compare foundational ethical theories (e.g., utilitarianism, deontology, virtue ethics) and professional engineering codes of ethics. They will be able to analyze how these frameworks apply to real-world engineering case studies.”

What They Did

All three tools produced a two-week content outline focused on addressing their respective first objectives. But their approaches varied—some stuck to the basics, while others took a more guided and structured path.

ChatGPT and Copilot created straightforward outlines based on the components of the objective. Each included weekly topics, suggested a short list of possible activities, and assessment ideas. These activities—like role plays, quizzes, comparisons, and reflections—were practical suggestions. ChatGPT went a step further and included a concise summary at the end, which I liked. It’s the kind of structure that caters to readers (and instructors) who want the outcomes.

Google Gemini, on the other hand, took a different route. It not only created a content outline but also added detailed guidance, essentially providing a mini-instructional design plan. The outline included learning flow, rationale, and even implementation suggestions. This could be extremely helpful for instructors with limited experience in curriculum design. Some might argue it overreaches by minimizing human decision-making, but isn’t that exactly what it feels like when working with an instructional designer? That said, the detailed nature of Gemini’s output might not appeal to those looking for something quicker or more high-level.

My Take

As someone with a background in instructional design and many years of teaching experience, I found Gemini’s approach the most pedagogically sound and helpful. It offered not just structure, but support. Still, the simplicity and clarity of ChatGPT and Copilot might better serve instructors who already know what they want and just need a little scaffolding.

Up next: how these tools handle creating in-class activities. Stay tuned.