Case-spiration: E23 - Real Inspiration

Success, 10 Questions and Failure using AI for Case Writing

Matthew Schonewille
September 16, 2024

Hey Fellow Case Writers!

I know that many of you are constantly wondering, can AI write cases or how good can those cases really be?

In this edition, I’m going to share with you 2 personal experiences and 10 questions I have had using AI for Case writing. They are connected to all of the work I have been doing learning what AI can do, how it works and what are the limits in AI Case Writing.

High-level, we need to be very aware that there is a new way to write cases. That new way might be better for some and worse for others. BUT, we can not ignore the fact that AI is here and writing cases is one topic that AI can do exceptionally well.

**** Side note, I am trying a new way to have the Audio of the Newsletter (not my voice), it will sound a bit different but let’s give it a try*****

Success Story Using AI

My first case exam writing experience with AI

One of my first experiments and successes with AI and Case writing was focused on using a Case Exam vs Case for discussion. I was teaching a marketing course and we had exhausted the cases that were in the text book. Students knew the cases and the answers were out there.

I offered to try and write a case for the exam that was written with AI but was focused on Marketing that was connected to our course. First, I had ChatGPT generate a story that was really interesting about Master Card Black - a credit card rewards plan. It had real information, figures, data and content that was realistic.

Then, I had ChatGPT align the case to the course learning outcomes and then to have specific elements from the course be part of the main flow in the case. I wanted environmental analysis, market research, content about STP - Segmentation, Targeting and positioning and lastly, have the case sound real.

ChatGPT did a really good job after all of the prompting and revisions. Overall, it took me about 60 minutes to write the case that made sense and was usable. I want you to digest that for a minute, 60 minutes for a NEW and Unique Case.

Then, the best part was having ChatGPT generate assessment questions. I had it generate assessment questions that were connected to the case as well as specific content in the course. This format ensured that if a student used AI to answer the questions, it would not know the specific context from our classes and if the student was clueless, they could only get some part marks.

Overall, this was a major success. I say success because:

I was able to create a unique case that didn’t exist anywhere
I created a new case that was interesting and aligned to my course directly
I was able to generate assessment content that was aligned to what I was assessing
I was able to adapt this case over and over again to fit my needs
I did not have to spend money on a case from a publisher
My class had not see this case and the assessment was authentic

So, the next question is, did it work? Yes, the case worked and students were able to do their exam that properly assessed what we were looking to do for the exam.

What was even better was that, the grades were right in line with the trends that we were expecting. That means that the case did not present easier or harder but right in-line with where it needed to be.

Failure Story Using AI

Case Writing about a topic I have no expertise in

Last year I was generating case studies for HR professionals/Students that were meant the be focused on common topics and content relevant to HR type classes. I am no HR expert but AI was here to help!…. or so I thought.

I generated a case on hiring practices and added in a racial bias in the hiring process. At the time the model I was using was ChatGPT 3.5 which was really bad a writing cases BUT I had no idea and that was what I had.

I initially generated a good starting point for the case that sounded realistic and compelling but I needed to make the case longer more then 600 words. I used a prompt called “Expand this section and add more details” which ChatGPT did but it did it so Badly. It added in filler and absolute crap. I initially thought, man this is good but when I really assessed and read the case, it was in major need of word smithing and revision.

Overall, in this scenario, simply asking the model to add more details or expand the case was not enough and resulted in a bad case and horribly written story that had no business being part of ANY class.

Now, the models are better and this would not go as badly. You still need to be more clear about what details you want more of or how you want the model to expand the case.

10 Common Questions and Answers

Below are 10 common questions that I have been asked or come across over the past 1.5 years learning about AI and Case Writing. It’s not an exhaustive list but should give you some context.

1. Can AI effectively generate case studies?

Yes, AI can generate case studies based on provided prompts or input data. It can simulate real-world scenarios by analyzing existing information, identifying key themes, and creating structured narratives. However, human oversight is necessary to ensure accuracy, context relevance, and ethical alignment with educational standards.

2. What are the limitations of AI in writing case studies?

AI-generated case studies may lack depth in nuanced analysis, critical thinking, and human empathy. They may not fully capture the complexities of real-world decision-making. AI may also inadvertently create content that is biased or not entirely accurate. This is why it's essential for educators to review AI-generated content for coherence, relevance, and quality.

3. How can AI help students understand case study methodology?

AI can be used to demonstrate various case study structures and writing styles, offering examples that students can analyze. AI can break down case studies into their core components, such as background, problems, solutions, and outcomes, helping students understand how to structure their own work.

4. Is AI-generated content original, or does it plagiarize from existing sources?

Modern AI tools generate unique content based on patterns learned from a wide array of texts rather than copying existing material. However, the risk of unintentional similarity to existing works is possible. It's advisable to use plagiarism detection tools to ensure originality when using AI for case study writing.

5. How can AI be used to support collaboration on case studies?

AI can facilitate collaborative learning by generating initial case study drafts that groups of students can refine. It can also help streamline feedback by summarizing key points and generating alternative perspectives, encouraging deeper discussion and teamwork.

6. What ethical concerns should educators consider when using AI to write case studies?

Ethical concerns include the potential for over-reliance on AI, misrepresentation of AI-generated work as original student thought, and the risk of bias in AI outputs. Educators should set clear guidelines about when and how AI tools are appropriate to use and ensure that AI serves as a supplement to student work, not a replacement.

7. How can AI help improve case study writing skills for students?

AI can act as a tutor by providing examples of well-structured case studies, suggesting improvements in grammar, coherence, and structure. It can help students practice writing by offering prompts and guiding them through the process of organizing and presenting complex information effectively.

8. Are AI-generated case studies suitable for all educational levels?

AI-generated case studies can be adjusted for different levels of complexity, making them suitable for various educational levels. However, educators need to tailor the output to the students’ cognitive and academic level, ensuring that the AI-generated content meets the learning objectives for each grade or discipline.

9. Can AI-generated case studies be used in exams or assessments?

AI-generated case studies can be used as part of formative assessments, where students analyze or critique an AI-produced case. However, when used in high-stakes testing, the use of AI should be transparent, and educators must ensure that students' own analytical skills are being evaluated.

10. How can AI aid in grading case studies?

AI can assist educators in grading case studies by providing quick feedback on structure, grammar, and basic analysis. While AI can offer preliminary evaluations, human review is crucial for assessing critical thinking, originality, and in-depth analysis

-------------------

Wrapping it all Up!

If you are interested in trying out the Free Case Study Prompt Generator .. Click the image below.

* Remember, please move this email from your promotions to your primary Inbox

If you have a podcast, I would love to speak on it -

Please reach out and lets start a conversation. [CLICK HERE]

Your Feedback Matters -

We love hearing from you! Share your thoughts, suggest topics, or case studies you'd like to see in future newsletters. [Feedback Here]

Thank you for being a part of our community. Together, we're shaping the future of education, one case study at a time!

AI In Education: I’m walking beside you in the weeds.

Matthew is the creator of the "Case-spiration," newsletter, a platform designed to share his extensive experiences and insights in case-based teaching from an educator's perspective. His primary goal is to empower faculty and staff in educational settings with the necessary tools and knowledge to excel in teaching and learning during this era of significant generational shifts. His approach emphasizes practical, case-based learning that prepares students for real-world challenges, fostering critical thinking and problem-solving skills via thought provoking scenarios.

Warm regards,

Matthew

Reply

or to participate.