The heart palpitations and hand-wringing in my academic field of Writing Studies began in early December 2022 when headlines such as “The College Essay Is Dead” in The Atlantic and “Freaking Out About ChatGPT: Artificial intelligence can crank out passable student essays in seconds. What are we going to do?” from Inside Higher Ed began appearing and being shared and discussed on numerous list-servs and social media groups.
Given our training as researchers as well as teachers, these initial Chicken Little “the sky is falling” posts were quickly followed by posts from folks who were testing the capabilities of ChatGPT by giving it actual assignment prompts from their classes or asking it to assess some of their own students’ writing.
During this initial fervor, I sought reactions from the group on campus to whom I often turn for richly-informed discussions about writing and the teaching of writing: the writing consultants. I began our December professional development meeting by asking the consultants if they had heard of ChatGPT, if they had tried using it and if any of the writers with whom they consulted had referred to it. A few consultants had heard of it, but none had tried using it, and they hadn’t had any writers bring it up in conversation. I proceeded to hand out an example of writing ChatGPT had generated in response to assignment prompts written by Ashley Squires, assistant professor of English and writing center director at Avila University. While the consultants agreed that the AI-generated writing was a bit “uncanny” in its mimicry of academic writing conventions, they quickly turned to talking about how they would consult a human writer of the text — a text that might be “passable,” but did not sufficiently fulfill the conditions of the prompt and was generally vapid and voiceless.
After I expressed my appreciation of their careful reading and thoughtful feedback, I shared with them some of the comments ChatGPT had provided in response to one of Squires’s student’s writing, as well as her acknowledgment that while some of the AI’s suggestions were not good, some were similar to advice that she would have given the student.
As we continued our discussion, the consultants repeatedly returned to what they felt were the uniquely human values of producing written texts. Why, they asked, would they completely cede the work of generating writing to a bot when the purpose of their writing process is to learn about a topic, to think critically about it, and to express themselves to an audience of other human beings? And they all agreed that a bot couldn’t produce truly creative writing because it couldn’t draw from lived, embodied experiences. I smiled at the sincerity of their perplexity, and suggested it was likely why they had become writing consultants in the first place.
Of course, the writing consultants are not representative of all student writers, and the difference between producing writing and writing to learn is often overlooked. Here I agree with the Writing Across the Curriculum professional organization, whose statement on AI writing tools argues, “Writing to learn is an intellectual activity that is crucial to the cognitive and social development of learners and writers. This vital activity cannot be replaced by AI language generators.” Having an AI generate text may well be more efficient, but the process of learning that occurs through the act of writing is lost.
Similarly, professors may be lured by the efficiencies of AI-generated responses to student writing, but if they delegate this work to AI, they will lose out on the opportunity to learn from their students’ writing — one of my favorite aspects of teaching writing — and the students will lose out on receiving responses from real human readers, not just artificial editors.
However, I’m not naive enough to think that all or even most students and instructors will eschew the efficiencies of natural language processing AIs. So, during the first class meeting of my FS102 this semester — which happens to be titled “The Robots are Coming!” — I told the students that I was not going to attempt to prohibit their use of AI text generators. Instead, I was going to ask them to acknowledge any help they received from AI — including idea generation, sentence structure, source summaries or citations — and describe how they incorporated it into their writing. I also told them that I would not accept any work wholly produced by an AI.
Because I scaffold the writing instruction and assignments in my classes, rather than simply assigning writing and collecting and “marking up” the finished product, students move from generative invention and brainstorming through revisions of multiple drafts for which they receive oral and written feedback, consider their choices and create revision plans before turning in the draft that I assess. Doing so, I hope, helps the students see that much of the hard work of writing and critical and creative thinking happens in revision. And it’s that work of thinking and making informed choices that I’m most interested in engaging in with them.
Seven years ago, my science fiction book club read “The Word Exchange” by Alena Graedon, a dystopian novel set in the near future in which people have become so reliant on technology to provide them information and complete their thoughts that when an evil corporation creates a virus that inserts gibberish in place of actual words, people panic when a “babble pandemic” ensues. Back then, in 2016, autofill was mostly confined to text messages — it hadn’t become prevalent in word processing, and the canned suggested responses in email were not common. During our book club discussion, I speculated that as more people opted for the efficiency of autofill, our language might begin to flatten and become less expressive. Back then, it seemed unlikely. Now, not so much. As a teacher of writing, I will continue to try to inspire in students the desire to write to learn, and I will continue to respond to that writing as a human reader. I hope others will do the same.