AI Policy in my Creative Writing class
Real-world, Monday-morning teaching. You're invited to weigh in.
We’re in transition. (There’s your Between connection.) A semester ago, this would not have been my policy on AI. A year ago, all but part 1 would have been unthinkable. I’m seeking comments, especially from other CW teachers. How are you addressing the use of AI in your classroom?
(The course is a lower-division course in reading & writing short fiction.)
POLICY ON AI & PLAGIARISM
1. This course is designed to help students maintain academic integrity by structuring assignments that build on one another and by requiring that we start many assignments in class. Plagiarism, the practice of taking someone else’s work or ideas and passing them off as one’s own, according to my Merriam-Webster online dictionary, is a complex term that probably deserves its own unit. Many students might not realize it is possible to plagiarize unintentionally; it’s even possible to plagiarize yourself. Remember that it is never wrong to ask if you’re unsure. In any case, students should be aware of the university’s policy and know that I will fully enforce it. Academic integrity statement here.
2. Submitting work produced by AI and committing plagiarism are not necessarily synonymous acts; however, it is important to know and abide by the approved uses of AI in this class. Remember, instructors vary significantly on the topic; what is okay here might not be okay in another academic setting.
a. First, and most importantly, AI can be regarded as a form of assistive technology. In that spirit, you may use AI however you like to produce your creative written results for this class (stories and story building activities). This means you may not use AI for other work such as preparing for class discussion, peer annotations or reflections on your writing even though it might help, because that is too much like just asking AI to do your work for you.
b. If you do use AI, I expect you to state how you used it and identify the results. This means that in addition to the assignment itself, you must add a note about your process. For example, in our activity on creating a fictional character, you will produce a brainstorming list that might be considered the “raw data” you feed AI. In that case, make clear that’s what you did, share what result you got, and discuss the process you used to get from that result to your final product. Note that AI’s result and your final product should not be one and the same.
c. We can discuss and even practice the approach of AI with every story-building assignment, depending on the interest of the class. It is not less work than producing the writing yourself; it is different work. While it is still creative activity, it depends more heavily on how you engineer your prompts. You are still responsible for input and final product. And you will probably need to decide if the learning you are accomplishing is acceptable to you. For example, the creative initiative required to imagine a plausible, compelling fictional character will not only produce different written outcomes than AI, it will likely result in different learning outcomes as well. No one knows enough about AI yet to say for sure. Monitor your own learning and, if you like, share what you are noticing with us.
3. It is still possible to commit academic dishonesty with AI despite my policy on it. If you ask AI for ideas and results repeatedly without providing your own imaginative or intellectual input you are cheating, plain and simple. If you then make misleading statements about how you used it, that, too, is cheating. As I think I’ve made clear, I am not the AI police. You might just succeed in fooling us, anyway. But if the only risk you’re taking in college is the risk of getting caught, you already know what you want; you just don’t yet understand the cost.*
*this is the sort of line I used to be able to say; not sure I can get away with it anymore.
"c." is for graduate students, mostly the last section. Too complicated.
If it were me, I would have an assignment built upon AI and AT THE END OF THE COURSE. Let them know you want to work with "their" ideas/language/whathaveyou and then experiment. Don't give in just yet and introduce the dragon at the beginning.
For me the first question, for them, is "do you want to read an AI-generated story?" So far, everyone I've asked says no. So then I ask them why. The answers (so far) tend to be a little tortured :) but I think it comes down to the idea that stories are supposed to be communications among humans, based in our shared humanity -- and an amped up autocomplete lacks that. Personally I'd rather read dictionary entries (which at least were written by people) than AI-generated stories, IF I know they're AI generated. *Knowing* they are AI-generated makes me very impatient with them.
I did have a student last semester who experimented by putting an GPT-paragraph in his essay to see if anyone would notice, and while no one realized he hadn't written it, almost everyone suggested he cut it because it was dull, repetitive, etc.
I definitely want to know how things go with this approach! :)