Assigning Generative AI in English Composition

An image of a student sitting at a desk in an English composition class, interacting with a chatbot on their laptop.
OpenAI. (2024). A student in an English composition class [AI-generated image]. DALL-E. https://www.openai.com/dall-e

Written by: Linda McHenry

When I first taught English composition—thirty years ago—word-processing software was not exactly a new thing. Nevertheless, my students needed instruction on the finer points of Microsoft Word or WordPerfect or whatever we were using then. The fact is my students today need various levels of instruction depending on their familiarity with word-processing software. And now generative AI has arrived, and I have an obligation to teach my students ways to use this new tool to optimize their writing processes. It took me little time to take another step forward in composition and technology.

This spring, I began experimenting with my pedagogy. As part of their brainstorming and prewriting process, I gave my students the wide-open directive to “Ask ChatGPT for help with some aspect of this essay assignment.” I then asked them to discuss—in real time, while and immediately after engaging ChatGPT—the prompts they were giving the AI and the results they received. Some students asked the AI to help narrow the topic of their essay. Some asked it to generate an outline for their essay. Some went so far as to prompt the AI to write a draft of their essay. One student, in a humorous act of rebellion, set poor ChatGPT up to fail, asking it to explain how to play softball.

After everyone reported some part of their work to the class, I asked students to tell us if the results they received were helpful, meh, or unhelpful. Generally, they were not impressed with the AI’s production. When I asked for a show of hands from those who felt like working with ChatGPT helped them feel more confident about some aspect of the writing task, many students raised their hands. As chaotic, comical, and sometimes frustrating as this exercise was, the introduction to using AI seemed to be a valuable experience.

In subsequent class periods, I moved on to teaching students how to ethically cite generative AI in their essays, following ethical documentation guidelines established by the American Psychology Association. Timothy McAdoo (2024, Feb. 23) writes a concise piece for the APA Style blog. Additionally, the Modern Language Association has developed instructions for incorporating generative AI in students’ writing. 

Our students must learn ethical ways of utilizing technology, so I require students to use generative AI specifically during brainstorming and prewriting. As a conclusion to every essay assignment, I require my students to write a reflective response, describing their writing process, explaining how they wrote, revised, and edited their essays. I also require them to include in this reflective essay all prompts and the subsequent text from ChatGPT. One pattern that has emerged from these reflective writings is that students found the AI to be inadequate for generating raw text. However, in those moments when they felt blocked or stuck, the AI was helpful in finding a way through the problem or a way back from the edge.

As is often the case, my students consistently surprise me with their candor in their reflective responses. Several of my ENG 101 students wrote that they found value in generative AI assisting them in selecting songs for a playlist essay. I had not expected them to need help with the song-selection task, but several indicated that they asked the AI to generate lists of songs for them. The students then chose their playlist songs from the AI’s list, writing that they chose the songs for their essay because “They are songs I know” and “These songs fit the genres for the assignment.” One part of the writing task, “Choose a handful of meaningful songs from among the thousands you have heard,” seems to have induced a degree of decision fatigue that I had not expected. But these students navigated the difficulty by asking the AI to narrow the scope of choices down to a manageable list.

Many of my ENG 102 students indicate they find value in asking generative AI questions about assignments that that they are unwilling to ask their instructors. They turn to AI when they “just want to know more without asking the professor.” Initially, I found this distressing. Students should turn to us when they lack clarity about our assignments. Explaining assignments is part of what we do as educators. But for some of our more anxious and neurodivergent students, such conversations can be difficult and unproductive. Here generative AI can be at least somewhat helpful if the alternative is that students have no direction at all.

Years ago, educators in my field of composition and rhetoric lamented the loss of penmanship while marveling at how polished typewritten texts appeared. Today, we face an equally transformative technology. If we delay engaging it until we fully grasp its possibilities and implications, we will have waited too long. I have decided to jump in, explore, and experiment. Please share your experiments, successful or otherwise, with generative AI in your courses.

Be the first to comment

Leave a Reply

Your email address will not be published.


*


This site uses Akismet to reduce spam. Learn how your comment data is processed.