5 prompts that explain how a writing professor flipped the script on AI

Assistant Professor Meagan Malone, Ph.D., teaches First-Year Composition and 300-level Professional Writing courses in the Department of English.

Prompt: You are a relatively new faculty member in the Department of English who specializes in teaching writing. In November 2022, OpenAI releases ChatGPT. Within weeks, star New York Times columnist Frank Bruni is asking “Will ChatGPT make me irrelevant?” and the Chronicle of Higher Education, in a piece called “AI and the Future of Undergraduate Writing,” leads with “Is the college essay dead?” A feeling of dread descends on — basically everyone. How should you respond?

Output: “I had a panic existential crisis in April,” said Meagan Malone, Ph.D., assistant professor in the UAB Department of English. “It was the vortex of despair. I went into [department chair] Alison Chapman’s office and said, ‘We’re not going to have jobs.’”

Flash forward six months and Malone is giving campuswide talks on teaching with generative AI [the technology behind ChatGPT and a host of competitors, including Google’s Bard, Anthropic’s Claude and others], offering classroom-tested advice to other faculty. And she is looking forward to the upcoming conferences in her field — rhetoric and composition, or “rhet/comp” — where she can hear what colleagues at other institutions have done. With nearly a semester’s worth of experience teaching two writing courses, the November 2023 version of Malone can offer April 2023 Malone a reasoned take: “It’s not that dire.”

Prompt: It is summer 2023. You are teaching two writing courses in the fall: First-Year Composition (EH 101) and Professional Writing (EH 315). All of your students can access AI writing tools for free. What do you do?

Output: In May, Chapman asked Malone and three other faculty members — instructor Amy Cates and associate professors Lilian Mina, Ph.D., and Danny Siegal, Ph.D. — to form an AI task force for the department. They searched and read and emailed and listened to podcasts and attended conferences, meeting repeatedly over the summer to compare notes and ideas on how to revise their courses.

Malone does not consider herself a tech fan. But “I like being relevant, and I like helping students,” she said. So she revised her courses extensively. “It is a lot of work; there is no getting around that,” Malone said. “Every single assignment I have meaningfully transformed. But I have learned a lot, and I already know how I will do it differently next semester.”

Prompt: Which generative AI assignments worked well, and which ones did not?

Output: Malone began her First-Year Composition course with several class sessions that covered the basics of generative AI and included debate-style discussion about the purpose of writing instruction. Students then wrote a piece explaining to Malone and their classmates their own thoughts on writing and the value of AI writing help.

Another in-class activity for the First-Year Composition course, which Malone called ChatGPT Playtime, was inspired by a talk that she heard at the Council of Writing Program Administrators conference. “The idea was for students to explore how we keep our own voices in a world of generative AI,” Malone said. Students prompted a generative AI model (they could use one of several listed by Malone, including ChatGPT and Bard) to write their biographies, seeding it with some details of their background. In one student’s case, based on her name and a few facts, the model “generated a very stereotypical view of her that was completely different from reality,” Malone said. “The other students were irate. It was a good lesson on how generative AI is not a source of truth.”

For her Professional Writing course, Malone recorded a video interview with a UAB English alumna now working in communications at a bank in Alabama, who explained how she and her colleagues (encouraged by their employer) were already using generative AI in their work. That interview sparked one of Malone’s assignment ideas. She charged her students, playing the role of employees at a small Southern bank, with developing an email response to worried investors questioning the bank’s stability in the wake of the (real-life) Silicon Valley Bank closure crisis in spring 2023. Students had to research the details of the Silicon Valley Bank crisis and how that situation was unique to its particular clientele and business model. They could use generative AI if they chose (to help them understand the details of the case, or to offer ideas on a structure for the email) but had to provide screenshots of all prompts, do a factcheck of its output and provide other documentation.

“Some students said, ‘I’m not going to do all that,’ and did everything themselves,” Malone said. “Others you could tell used gen AI without giving it much thought. But the ones who did the best, surprisingly, were the students who asked for AI help in structuring the email, and then built on that with their own words and research.”

In another assignment for the Professional Writing course, the students were given three blocks of AI-generated text by Malone and charged with revising and incorporating them into a single new paragraph using the techniques that they had learned so far in the class. “This activity did not work out as well,” Malone said. “The students seemed to be dazzled by the output from gen AI” and were hesitant to change much of anything, despite cluttered sentences and poor organization.

Prompt: Do you think that generative AI means the end of writing instruction?

Output: “Changes were coming to writing instruction regardless,” Malone said. “Gen AI is just accelerating a movement that has been going on for the past 10 years.” In fact, UAB was already in the middle of a curriculum revision for its First-Year Composition classes, she explains. “The philosophy is that assessments should be geared to demonstrating process and assessing student reflections along with the drafts and final product,” Malone said. “That was always going to be more labor. Gen AI is pushing us faster into a change we knew we needed to make anyway.”

Prompt: You are talking with a faculty member thinking about how to add generative AI into their course. What is your best piece of advice?

Output: “If there is anything that all people should communicate to their students, it is this: Generative AI is not a truth-generating machine,” Malone said. “Whether they had used it before or it was all new to them, none of my students knew that it is not a search engine. When I told them it was a next-word prediction algorithm, that it is not drawing upon a bank of facts, they were surprised. If it does say a truth, it is serendipitous. It is trained to give you what it thinks you want to hear. Students need to understand that.”