Writing Case Studies Using Generative AI: Intimate Debate Case Study 

Created with Firefly: Cartoon woman with headphones and robot next to her at a computer

This article is a two-part series. Click here to read article 2, Writing Case Studies Using Generative AI: Interactive Role Play.

Case studies consist of a short narrative followed by guiding questions that help learners think through the steps of solving a real-world problem. Cases are engaging, ask learners to work at higher levels of Bloom’s taxonomy (i.e., apply, analyze, evaluate, create), and have been shown to be an effective way of learning (Lee et al., 2009; Thistlethwaite et al., 2012). 

Generative AI (GenAI) tools such as ChatGPT make it easier and faster for educators to write case studies. These tools can serve as a research and writing assistant. They can also create unique, interactive, role-playing experiences that transform how cases are used in the classroom. This series of two articles will describe ways to use GenAI tools to create case studies.  

  • The first article (this one) will provide instructions for creating an intimate debate case study (a case study where students evaluate evidence for two sides of a controversy). Crucially, this method of writing a case will allow the educator to stay in charge at every step in the design, and it will use GenAI as a research and creative assistant. 
  • The second article will introduce a novel type of case study, an interactive role play. This creates a new case study each time students enter the prompt the educator assigns into a GenAI tool. The case is adaptive with the GenAI going in different directions based on the students’ responses. It’s a personalized, adaptive instruction. The drawback is that educators cede control of the experience to the GenAI. 

The two articles provide easy-to-use, step-by-step instructions that novice GenAI users can try and that more advanced users can use as starting points for more elaborate case designs. 

What you need to know before you begin?

What is a case study?  

A case study is a short story followed by (or interspersed with) guiding questions that put students in the shoes of a key player in the narrative so they can practice the skills of thinking in that discipline. I am a scientist, so for me, cases are tools to help learners practice the skill of thinking like a scientist. Cases ask learners to formulate hypotheses, provide context for designing experiments, encourage them to predict the results of those experiments, and help them analyze authentic data. 

There are many types of cases and each vary in their structure (Herreid, 1998). Some examples include intimate debates, clicker cases, and progressive disclosure. You can read about them on the National Center for Case Study Teaching in Science website (see the section, Teaching with Cases). The instructions below explain how to use GenAI to write an intimate debate; however, GenAI could be used to write cases in any format. 

What is GenAI? 

Generative AI (GenAI) is a subset of artificial intelligence that uses machine learning models to create new, original content, such as images, text, or music, based on patterns and structures learned from existing data. Another way to describe them is as sophisticated chatbots, where a user engages in an iterative conversation. The noteworthy aspect of this conversation is that every GenAI response is unique. In other words, if you ask the GenAI a question, it will likely give you a different response every time you use it. If you can Google a response, then GenAI is not the right tool. It’s best used for original responses that combine information in novel ways. 

What are some of the considerations when using GenAI? 

There are many thorny issues that come with using GenAI (Brown et al., 2023; Trust, 2023; Michel-Villarreal et al., 2023). The first is intellectual property since the data that the GenAI tools were trained on are not attributed (and indeed, many authors did not consent to their work being used in this manner). There are issues of academic integrity and plagiarism for users (like students) who may offload their work to the tool.1 Also, entering private or proprietary data into the GenAI tool may invertedly compromise the integrity of that data, breaching privacy and data security. There are also issues related to environmental costs, social justice, and labor practices. Finally, there is the issue of GenAI hallucination. The workings of GenAI is such that, when it does not know something, it makes it up! For this reason, the best users of GenAI are those who have existing knowledge about the topic so they can critically review and evaluate the trustworthiness of the output. This is often an issue with learners using GenAI – they do not yet have the expertise required to provide this guardrail. 

What GenAI tools can you use?  

ChatGPT is perhaps the most famous GenAI tool, but it is by no means the only one. Below are a list of tools, with more becoming available all the time. The following are free but require registration: ChatGPT (model 3.5 is free, and the model 4.0 requires a paid subscription), Microsoft Edge offers Bing in Creative Mode (also called Copilot), Google Bard (a GenAI tool that gives very different output), and Anthropic’s Claude.  

Some tools that do not require registration include You.com, TalkAI, and Perplexity.ai; however, they do usually contain advertisements as a way to support their product. 

Citing GenAI 

Many educators want to demonstrate best practices in citing sources for their work, including citing the GenAI tool that they used in the creation of their case study. Citation standards are emerging and many of the citation styles have come up with guidelines for citing the use of such tools. 

Since GenAI generates unique responses every time it is used, its output is not reproducible. This poses a challenge since an important reason to cite work in scholarly practice is to allow others to consult the original work. Therefore, one approach to citing GenAI work has been to save the chat (most GenAI tools allow users to do this) and provide a link to this original interaction as part of the citation. This is demonstrated below. 

Method 1: Research assistant for intimate debate 

In this first method, educators stay in control of the case during its design. They use the GenAI tool as a creative, research assistant. Educators guide it, repeatedly refining the tool’s output, and double-check everything it says for accuracy. Educators do not cede control of the design process to the GenAI. It’s like working with a student assistant, where the educator provides tasks to research or brainstorm. The assistant completes the tasks and comes back to the educator with their findings. The educator then reviews the work, chooses an option that most aligns with their vision, and progresses from there. 

Using this approach, GenAI can be used to create any type of case study, but as an example, we will create an intimate debate (Herreid & DeRei, 2007). An intimate debate is a short story about a real-world controversy—a complex situation where the course of action is unclear. There are two competing approaches to solving the situation, and each has its strengths and drawbacks.  

In teams of four, students read the background scenario. Then, they split into teams of two and each read a bullet point list of evidence that could be used to support one side of the controversy (i.e., each bullet point will either summarize the data, provide results of the study, or contain facts pertinent to the situation; we will call them the Pro handout and the Con handout). Students review this evidence and formulate arguments to support their side of the controversy. Students then reform their original group of four (and the data sheets are taken away, though students can keep their own notes containing the arguments they developed from the bullet point data) and take turns sharing what they learned. As a team, the four students examine the two sides and the evidence supporting them. They then decide on their preferred course of action about the controversy and articulate how they came to this decision. The following figure illustrates the steps of an intimate debate. 

An example of an intimate debate case study is the case of golden rice (Prud’homme-Généreux, 2009), which is a genetically modified rice. The staple crop has been engineered to synthesize vitamin A, one of the most common nutritional deficiencies in the world. Despite the GMO being free of charge for subsistence farmers for decades, it has still not been adopted widely. Why? What are the arguments for and against its use and should it be used more widely? 

To create an intimate debate, an educator needs to develop three resources (indicated with red checkmarks in the figure above showing the steps of an intimate debate): 

  1. A short narrative that introduces the real-world scenario and the dilemma, which then invites students to make a decision about how to solve this situation. 
  2. One page containing bullet points that summarize the evidence for one side of the argument (the Pro handout). 
  3. One page containing bullet points that summarize the evidence for the other side of the argument (the Con handout). 

      The GenAI tool will be used to accomplish four tasks in designing this case study: 

      1. Identify a controversy related to your learning outcome
      2. Write the introductory paragraph
      3. Research and write the Pro handout
      4. Research and write the Con handout

            Brainstorm controversies 

            I am an instructor teaching [biology] in [first year of university]. I want to use a real-world controversy related to the topic of [mitochondria] to help my students learn more about [cell biology and the interplay of science and society]. I would like you to suggest 5 possible real-world controversies that I could use to achieve these goals. Cite your sources and do not create them. 

            Examine the output. You may not get great results on the first input. Remember that a conversation with a GenAI is an iterative conversation. Your interactions progress as you refine the output by asking the tool to tweak what it has produced towards a product that you want (e.g., you may tell it to focus on controversies that have taken place in the United States and in the past five years). 

            Note that the direction: “Cite your sources and do not create them” works better in some GenAI tools than others (e.g., it works better in ChatGPT 4.0 than ChatGPT 3.5), so remain vigilant and check the proffered sources for validity and accuracy. You can see a sample output from Chat GPT 4.0 for this prompt. This is an example of a saved interaction with a GenAI that can be used in a citation. 

            If you decide to write your own prompt, be sure to specify: 

            • Your contextthe level/grade of the students and course/discipline 
            • The taskbrainstorm real-world controversies to teach a topic 
            • The contingencies: limit to five ideas and do not invent sources (note GPT 4.0 will act on this request; GPT 3.5 may not) 

            Introductory paragraph 

            Select one of the controversies suggested in the above step. Your next task is to create the introduction to the case. The introduction will set the scene, explain the controversy, and assign the task to learners (review the evidence, pick a side, and explain why that side was chosen). Here is a proposed prompt. As above, please replace the sections in [BOLD] with information from your classroom context. 

            I am an instructor who teaches a [second year nutrition] course to [university] students. I would like to use the controversy over [the benefits and risks of a ketogenic diet on health] to help students [evaluate contradictory information on a health outcome]. I am looking for a short introductory to this controversy that I could give students to introduce this as a case study. Write this introduction. It should be [no more than 300 words]. The description should explain the premise of the controversy and end with a clear call for students to evaluate the information and make a decision about their position on this controversy. 

            Remember that if the GenAI doesn’t yield something that is suitable, you can refine its output by pointing it towards what you want. Here’s the prompt I used to refine the output: 

            This is excellent. However, I would like you to use a real-world example as a trigger for the case study. Can you re-write this and this time include allusion to the real-world situation as a reason for exploring [the benefits and drawbacks of a ketogenic diet on health]? 

            Other suggested ways to refine the output include: Shorten the text you have created to 200 words maximum; or write the same information but this time using an informal writing style with some humor; or be sure to add analogies when you describe X in the paragraph. 

            You can see a sample output from Chat GPT 4.0 for this prompt. In this example, the GenAI initially gave a generic narrative to introduce the need to investigate the health benefits of a ketogenic diet. However, after more refinement, it changed the introduction to refer to the actor Chris Hemsworth and his dietary choices as a hook for the case study. This is much more engaging. 

            Pro and Con handouts 

            The heart of an intimate debate case are the Pro and Con handouts. The GenAI tool will search for, compile, and write a list of evidence that could support each side of the argument (one side per handout). Here are the proposed prompts to generate these handouts. As for the previous prompts, you will want to modify the text to your context. 

            Pro handout 

            I am writing an intimate debate case study that will encourage my students in [a second-year course] on [nutrition] to [evaluate the benefits and drawbacks of a ketogenic diet on health]. As part of this case study, I plan to give one group of students a sheet of information containing 10 bullet points that support the idea that [the ketogenic diet is a healthy choice]. Create this list of 10 bullet points. Each bullet should list a verifiable fact, data, [or results from an experiment] or study that could be used to support the position that [a ketogenic diet is healthy]. Cite your sources and do not invent sources. 

            Con handout 

            Continue the above conversation with the following prompt:
            This is excellent. Now can you create a similar ten bullet point list, but this time, select facts, data, and [the results of experiments] or studies that focus on [the drawbacks of a ketogenic diet on health]. Again, cite your source and do not invent sources. 

            You can see a sample output from Chat GPT 4.0 for this prompt. In these prompts, 10 bullet points are requested. Ten bullet points are ideal for a 50-minute class time with undergraduate learners. You may, of course, customize this number to suit your pedagogical context. Note that you may also request more than 10 bullet points from the GenAI, and then select the 10 best ones for the student handout. 

            For this output, it is particularly important to double-check the citations. Make sure they exist, that the information described in each bullet point can be traced to the source cited, and that it is accurate. You may also modify the prompt to ask the GenAI to research data from specific sources, such as government websites, scholarly publications, or newspapers, etc.

            Putting together the case 

            Voila! You now have all the pieces required for an intimate debate case study. Instead of taking days, weeks, or even months to identify a good controversy, writing an introduction, and researching evidence to support each position, the whole process can take minutes. The bulk of the work is in directing the GenAI and in verifying its sources. 

            In the next article, we will examine how GenAI can be used to create an interactive role play. Here, the GenAI doesn’t just create the case, it also adapts to each student’s response, taking the case in a different direction depending on student choices and level of understanding. It’s an emerging way to use GenAI that provides customized education at a time when our classrooms are becoming larger and where we cannot tutor each student one-on-one.


            Dr. Annie Prud’homme-Généreux is an educational developer and instructor at the University of British Columbia. She is completing a master of online education and has designed and facilitated several workshops to help educators incorporate GenAI tools into their practice. This includes the OER faculty development course Forward Facing Assessments, downloadable free of charge from BCcampus. 

            References 

            Brown, H., Crawford, S., Miffitt, K., Mendolia-Moore, T., Venins, D., & Weiss, J. (2023). 7 Things your should know about generative AI. EDUCAUSE Review. https://er.educause.edu/articles/2023/12/7-things-you-should-know-about-generative-ai   

            Herreid, C. F. (1998). Sorting potatoes for Miss Bonner: Bringing order to case-study methodology through a classification scheme. Journal of College Science teaching, 27(4), 236-239. https://static.nsta.org/case_study_docs/resources/Sorting_Potatoes.pdf  

            Herreid, C. F., & DeRei, K. (2007). Intimate debate technique: Medicinal use of marijuana. Journal of College Science Teaching, 36(4), 10-13. https://static.nsta.org/case_study_docs/resources/Intimate_Debate_Technique.pdf  

            Lee, S. H., Lee, J., Liu, X., Bonk, C. J., & Magjuka, R. J. (2009). A review of case-based learning practices in an online MBA program: A program-level case study. Journal of Educational Technology & Society, 12(3), 178-190. 

             
            Michel-Villarreal, R., Vilalta-Perdomo, E., Salinas-Navarro, D. E., Thierry-Aguilera, R., & Gerardou, F. S. (2023). Challenges and opportunities of generative AI for higher education as explained by ChatGPT. Education Sciences, 13(9), 856. https://doi.org/10.3390/educsci13090856  

            Prud’homme-Généreux, A. (2009). Golden Rice: An Intimate Debate. The NCCSTS Case Collection. https://www.nsta.org/ncss-case-study/golden-rice  

            Thistlethwaite, J. E., Davies, D., Ekeocha, S., Kidd, J. M., MacDougall, C., Matthews, P., … & Clay, D. (2012). The effectiveness of case-based learning in health professional education. A BEME systematic review: BEME Guide No. 23. Medical Teacher, 34(6), e421-e444. https://doi.org/10.3109/0142159X.2012.680939  

            Trust, T. (2023). Essential considerations for addressing the possibility of AI-driven cheating, Part 1 and 2. Faculty Focus. https://qa.facultyfocus.com/articles/teaching-with-technology-articles/essential-considerations-for-addressing-the-possibility-of-ai-driven-cheating-part-1/