The typewriter didn’t “end” handwriting. Cliffnotes didn’t “end” reading. Wikipedia didn’t “end” research. Smartphones didn’t “end” community building. Zoom accommodation didn’t “end” interpersonal collegiality. In that same vein, I’m struggling to grasp what exactly it is that’s “ending” with generative AI in the classroom. From casual conversations with colleagues, it seems we’re experiencing some growing pains with the new technology that is now inevitably a part of our lives as educators, and not without reason. This is, for better or worse, forcing us to shift perspectives and practices as they relate to both our learning goals and their execution. The consternation between resisting the use of generative AI or incorporating it into our pedagogy is the current topic of conversation, but rather than fight change, I’m trying to go inward and use this opportunity to grow as an educator and test my own capacity to be creative in my course design. It’s easier said than done for sure, but I also think we’re all allowed to show ourselves some grace while we figure this out.
We tend to lose track of time, and before we know it, fifteen years have passed and the learning tools we have depended on for years have become a bit dusty or, dare I say, outdated. We are amid one of those moments, and just like we pivoted and incorporated new techniques in our teaching to account for online course construction during the height of the pandemic, I am trying my best to embrace generative AI and look at this as an opportunity for me to re-center my students and the context in which they live at the center of my pedagogical worldview. Students experience college differently than I did two decades ago, and it’s important to remember that when I construct my courses. The moment I feel that how I’m comfortable teaching is the only acceptable way to build and test student skills signals to me that it’s time to examine my approach and become comfortable with new teaching and learning strategies.
That said, one of the Women’s and Gender Studies Department’s core lower-division courses, WOS225: Intro to Women’s and Gender Studies, has a writing designation, so the enhancement, testing, and practice of writing skills are a keystone in the makeup of the course. Generative AI tools like ChatGPT pose a fundamental challenge to classic methods of building the relationship between critical thinking and writing. I took this as an opportunity to break down what the core values of the course were along with the main skills we wanted built into a writing course. Ultimately, key concepts like intersectionality, agency, and structural inequality, among others, are critical to be explored, and learning to practice implementing those concepts in student writing through reflection and synergy are key learning goals for the writing process. The traditional “incorporate these terms from the readings into a 1000-word essay” becomes a bit more difficult to assign when generative AI can produce an essay using those terms in near indistinguishable ways from student writing, thus eliminating the skill building inherent in the assignment. So, why not utilize the very system of AI to have students practice those skills while also critiquing the limitations of generative AI in the process? Below, I will briefly describe the original construction of a key essay I assigned in the class historically, and how I’ve updated the assignment to incorporate generative AI into the assignment design itself while maintaining attendance to the desired learning outcomes.
The “Social Location Essay” has historically been a short autobiographical essay due at the end of the first unit of the semester (around week 4) that incorporates texts from the class to reflect on the ways that one’s identity impacts their worldview. Part of the assignment description reads:
Based on the introductory readings from class, you will write a short autobiographical essay. Citing 2 class texts, you will outline the ways that your social location (i.e., your gender identity, race, class, religion, sexuality, etc.) has impacted your outlook on life, politics, the concept of feminism, and the roles of the women in your life. This is due early in the semester, so it may very well be that you have questions, hesitations, and skepticism, but feel free to include these thoughts in your essay. The main goal here is for you to think about how the concept of “intersectionality” and your matrix of identity (social location) have impacted your life, beliefs, and experiences thus far. These should be roughly 800-word essays.
Upon initial reflection, I assumed that a “personal” essay would be difficult to recreate with generative AI, but it is surprisingly simple, given that we know very little about the personal lives of the students we teach, and with the right prompt, systems like ChatGPT can easily make a reflective essay including any number of theoretical frameworks. Students are even getting crafty and after creating essays with generative AI, running them through paraphrasing software which applies plain language and removes much of the scholarly jargon that is usually the telltale sign of an AI generated paper. A part of me admires the dedication to the bit.
With this in mind, I made this very element I was trying to avoid a key fixture in the new assignment. To start, I had everyone in the class open generative AI software together during class and insert an essay prompt that seemingly addressed their own intersections of identity. For instance, my example prompt to the class was,
“write a 500-word essay about the college life of a queer, gender nonconforming, Korean American, video gamer.”
The story that popped out was ridiculous, hilarious, and in no way described who I was or what my experiences were. We all had a great laugh as I read my wildly inaccurate AI short story to the class about my electric blue hair (this must be the gender nonconforming bit), my strange love of Disney World (the queer bit, obviously), and my comically stereotypical overbearing Korean mother (that’s just garden variety racism). My students put theirs in, including elements like gender, race, religion, and many included additional elements like “business major,” or “cheerleader,” or “aspiring teacher.” We shared a mixture of laughter and reflection from hearing the ridiculous and, at times, problematic stories that the software had created based on their identities. A Latine student that was born and raised in western PA was now, according to generative AI, born and raised in East LA. A straight cheerleader was now, according to generative AI, madly and secretly in love with another of her fellow cheerleaders (so much for honoring the prompt!). A Black student who is a dancer, according to generative AI, grew up in the glow of a diverse and stimulating dance community, when in reality the student shared the often-painful experiences of being the only Black woman in her dance environments throughout her entire life.
It is no secret that generative AI software can only enact these prompts based on the data that the software has been fed or has access to. That said, there have been numerous articles that have outlined the problematic nature of AI data sets. They’re often racist, sexist, embody cultural stereotypes and misconceptions, and are deeply Eurocentric, often drawing “impartially” from problematic data that already exists online. Have you ever seen the comments on a Reddit thread concerning sexism or racism? Oof. To think that that data could be feeding generative AI software is scary, but that kind of data is one of the many strands that feed it. So, the classroom activity became an opportunity to expose some of those truths, dialogue about them, and gave the class an opportunity to challenge some of the stereotypes that feed generative AI, allowing students to become the experts of their own experiences while also applying a concept like intersectionality in real time. The exercise was a success, and students were actively engaging with the software in a way that was personal, vulnerable, and in community with one another.
The writing assignment, then, was to take that 500-word essay that was created, and directly address it and reflect on it. The idea was to identify what was wrong about the story and to use course texts to explain how concepts like intersectionality are so profound because they give us a way to intellectualize how our own elements of identity work together to create our unique and beautiful lived experiences. Part of the assignment description reads as follows:
Citing 2 class texts that discuss the concept of intersectionality, and in response to the generative AI story that we created in class, you will outline the ways that your social location has impacted your outlook on life, your politics, concepts of feminism, and the roles of the women in your life. You should include your generative AI story at the end/or beginning of your essay (clearly labeled), and you should be directly referencing it throughout your essay (e.g. what did it get right, what did it get wrong, and what does that make you think about how we as a society make grand assumptions about entire groups of people based on specific aspects of their identity). The main goal here is for you to think about how intersectionality and your social location have impacted your life, beliefs, and experiences thus far.
Asking students to use the generative AI software in advance and to respond to the product it created enabled them to feel personally compelled to take their own positions, because their own life stories were at stake. The essays the students responded with were beautifully reflective, honest, and in my opinion displayed a much more nuanced use of course terminology than the assignment had in the past. The semester that I implemented it, class grade averages on the assignment were significantly higher than they were in past years. I believe it is because the students not only felt more was at stake, but through the activity they actually gained a deeper understanding of intersectionality to apply for the rest of the semester. It was personal and it was utilizing the technological context in which they currently live and engage.
In a way I’m thankful for the challenge that generative AI software has provided for me. I was content to create assignments that reflected the way I had always been taught to think about writing mechanics, theoretical application, and critical thinking, but this challenged me to think more deeply about context, think more deeply about process, and to practice a form of generous engagement with my students that enabled us to share vulnerable and generative space with one another. I hope this can serve as some inspiration for my colleagues to rise up to the challenge of thinking about generative AI from another vantage point. We all become better educators when we center our students and the world they live in, so why not use this as an opportunity to do just that?
