Generative AI
“It’s a matter of shifting your mindset from ‘what do I think of it?’ to ‘what else could I think of it?’ How else could I approach this? This sort of thinking transforms AI from dead weight, to a sail, and then ultimately to an Iron Man suit. Your imagination is the only limitation, but imagination is a function of exposure.”
I have found generative AI to be a fascinating and versatile tool, particularly for ideation and creativity. I most frequently use it to “talk out” ideas or work through mental logjams, but I have also started a creative side-project using AI art to illustrate a fictional narrative told through Instagram posts.
The sample lesson plan to the right was created during a training session, and while I have taught from it, the artifacts below include more representative examples of my ideation process.
Ideation and Generative AI Resources
Grant, Adam. Think Again. New York: Viking, 2021.
Mollick, Ethan. Co-Intelligence: Living and Working with AI. New York: Portfolio/Penguin, 2024.
Utley, Jeremy, Perry Klebahn, and David Kelley. Ideaflow: The Only Metric that Matters. New York: Portfolio, 2022.
Utley, Jeremy, and Hendrik Werdelin. Beyond the Prompt. (podcast). https://www.beyondtheprompt.ai/
AI Roundtables
AI works better when it knows who we are and who it is. We can accomplish the first goal by leveraging the Custom Instructions: I provided general information about myself, but also my top five motivations from TruMotivate, top five strengths from Gallup and notes about my neurodiversity and mood disorder. To accomplish the second goal, I have formed several roundtable groups of experts from a variety of domains with whom to discuss my various projects in progress.
It’s always a good idea to be sure that the AI understands the prompts (in fact, I frequently end my prompts by asking, “Do you understand?” or “What additional context do you need?” or simply “What questions do you have?” In the case of my my expert roundtables, I have the AI characters introduce themselves. These discussion have proven exceptionally useful in talking out ideas, and while the AI responses are frequently rather pedestrian, this actually works in my favor—I’m good at getting the gist of things, but frequently lose track of the details.
At first, I drew all the names for my experts from the The Court Jester, an old Danny Kaye movie, but I realized this might be teaching the AI a white-male bias, and so recently I have been deliberately including such non-Western names as Watanabe (a chronobiologist) and Chhuon (a Vice President of Research and Planning at a Liberal Arts college)
Appropriate Usage of Generative AI
I am convinced that students cheat and plagiarize because we teach them to. The combination of learned helplessness and needlessly high-stakes assignments, particularly in writing, convinces them that it is simply too risky to attempt anything on their own. Accordingly, it is unsurprising that, given a tool that can write more-or-less coherent essays, it would prove tempting for students to take shortcuts—particularly later in the semester for students suffering from anxiety disorders. The results are hilariously obvious and quite awful.
I see this as an opportunity, however: my writing pedagogy in DIS 1000 is focused entirely on undoing the damage wrought by poor writing instruction, and since ChatGPT excels as a thinking partner, I began using it in class to demonstrate effective ideation. The examples on the right used the prompt from the Religious Music project, but asked for assistance in thinking through the experience, rather than writing it up. I entered this prompt in class, and we talked through the AI’s responses in real time. While not all of the questions it asked were useful (the second is particularly poor), the good ones were very good, and effectively illustrated the kind of thinking I would like to see in the projects.
I strongly encourage my students to use the Writing Center on campus: I have the program director come in to speak to DIS 1000 every semester, and offer additional pass credit for students who follow up in the “Connection to Your Major project,” our capstone writing piece. I also took advantage of some particularly poor AI-generated drafts from previous semesters to use as classroom examples. After the students worked through what makes them problematic, we fed them back into ChatGPT to hear its critique.
DIS 1000 Music, Resilience, and Life Skills: Interactive Quiz
For several semesters, I had been dissatisfied with the students’ ability to define and differentiate the elements of music (timbre, texture, rhythm, melody, and harmony). We addressed each of them in a series of creative or analytical projects, but by the summative song analysis, I wasn’t seeing the fluency I expected. I decided to reinstate formative quizzes, but I wanted them to be both silly, to defuse any test anxiety, and also interactive, to provide immediate feedback for wrong answers. I thought it would be interesting to use a structure like the old “Choose Your Own Adventure” books, and built the basic structure with MS Forms.
ChatGPT writes in a sickening flowery style, which I thought would be perfect for creating an over-the-top adolescent adventure, but also provide several opportunities to demonstrate that 1) ChatGPT writes terrible essays, and 2) everyone can tell when you’re using it. I worked with ChatGPT to translate the concepts into a haunted house narrative, and as a bonus, illustrated it with AI art from CivetAI. The result was surprisingly effective: many students answered every question correctly on the first try, and while I still need to work on getting them to the Analyze level of Bloom’s Taxonomy, I will keep these quizzes in the syllabus.
HNR 3603 Music and Identity: Vignoles Challenge
My goal with this assignment was to encourage a deeper engagement with the first major reading of the course, to strengthen the students’ understanding of the four aspects of identity: Individual, Relational, Collective, and Material, and encourage divergent thinking. I had the idea of incorporating the mechanics of the game Smash-Up, and used ChatGPT as a thinking partner in order to work out the details. The finished assignment was as follows:
Challenge: interrogate the an aspect of identity as assigned, develop deeper insights, and communicate them to the class metaphorically with custom-designed Smash Up decks
Constraints: Must use the HPU library website, work in six-person teams, and complete the work in four days
Criteria: to succeed in the challenge:
Pull at least three articles from the VIGNOLES ET AL. bibliography
Read, annotate, and discuss the ideas (be able to show your work)
Synthesize the ideas into a (creative) 20-card faction deck with ability and action cards, consistent with the game mechanics (minion abilities consistent with their faction)
Present the deck
One of the students expressed some trepidation about making the cards, which was something I hadn’t considered. Another member of her group simply shrugged and said “Use an AI art generator”—a happy accident that unleashed considerable creativity and returned some hilarious results. From that point, it was an unspoken assumption that every assignment would incorporate AI art illustrations, including the final course website.
Sample student cards
