Teaching the Next Generation of Writers in the Age of AI

As a writer living in a tech city, the dread of AI has grown unavoidable despite my best efforts to fully avoid it. Layoffs in the tens of thousands are hitting the companies that are the backbone of the Seattle economy, or threatening to, while I worry whether I (and everyone else) will be replaced by robots at any moment. I’m constantly looking around rooms and wondering “whose side are you on?” — then waiting for my stomach to drop every time a friend or family member casually mentions using ChatGPT or Claude.

It feels weird admitting that the only thing I use AI for is initial interview transcription, before going through with a human eye and ear afterwards. Wherever possible, I try to turn AI features off, because they're often misleading or incorrect, and whenever they try to write or summarize for me I can feel my writer brain clicking down a notch. 

Perhaps this has been to my detriment in some ways. My business journalist husband told me recently that there are some AI tools that could, conceivably, help me with the large amount of research I do for my book projects. These could enable me to search faster through various resources without painstakingly reading every out-of-print book or academic paper on the subject. I have staunchly resisted even dipping in a toe, because I feel — like many readers of Scratch — deeply attached to writing and even researching as an art form, whether the outcome is journalism, or literature, or something in between. I worry if we start outsourcing any part of our thinking, and analyzing, and yes, communicating, to a machine, we won’t be able to actually reach other humans with our work. At least not in the way we’re hoping to. 

As an artist, I would also like to think that what I do matters, and I am increasingly faced with the existential possibility that many people in the world disagree. 

My friends and colleagues in academia have not been lucky enough to completely ignore this technology. Schools everywhere went from policing AI use to encouraging their students to use it all the time, while also firing people for using it too much. Meanwhile, educators still have the hardest job of all: preparing the next generation for a very uncertain future while being pulled in different directions. 

Recently, I took my most burning questions about AI in scholastic writing to five writing professors in higher ed and asked them how they're talking to students about this technology — and what they’re hearing in return.

The following interviews have been lightly edited for length and clarity.

Note: This post contains Bookshop.org affiliate links. Purchases made through them directly support independent bookshops and help keep Scratch running.

Daniel Cryer

Associate Professor of English at Johnson County Community College
Overland Park, KS
Writes: The AI Minimalist (Substack)
Courses Taught: Composition, Technical Writing, Writing the Environment 

Do you have an AI policy in your creative writing or composition courses? Is it anything other than “No AI allowed”?

I teach composition but not creative writing, and I do have a policy that is (necessarily, I think) more complicated than “No AI allowed.” It takes up nearly a full page of my syllabus, but the short version is: Everything in your work not clearly marked as coming from somewhere other than your own brain needs to be yours, and if it isn’t, that’s plagiarism. This includes ideas, too, not just direct quotation. In other words, taking something ChatGPT produced and putting it into your own phrasing isn’t “yours,” just like doing that with a human source isn’t yours.

But this still doesn’t cover all possible uses. So, to cover other uses I ask students to include a “technology acknowledgment” section in their papers that lists what assistive or generative text they used: spelling and grammar checkers; AI for research, feedback, or editing; translators for ESL students, etc. And I spend some time in the policy discussing why it’s generally a bad idea for students in the early stages of their education (I teach at a community college) to use AI at all.

For many reasons, that’s not a realistic expectation. [AI is] increasingly a feature of their writing environment, and of course that’s by design, with students seen as a customer base from a very young age. 

I think we are at the point where not using it is swimming against the tide — it takes more effort than using it. I will also say that in some of my classes, we talk about AI a lot. In my more advanced composition classes, one full unit is centered on an AI proposal argument, where students identify a problem AI is causing, or a problem AI is being used to solve, and they describe the problem and argue for a solution. In this unit, we do use AI some as a research tool, but we spend much more time investigating, critiquing, and discussing it.

How much have you shifted methodology in terms of teaching writing now that AI LLMs are ubiquitous?

A lot in some ways, not much in others. My classes are structured the same way they were before November 2022: two or three units, each leading to a major project with smaller scaffolding assignments along the way.

All of the big changes I’ve instituted have to do with getting familiar with their in-process work before I see polished finished products. The first big shift I’ve made (along with many other teachers) is assigning much more in-class writing. For the first half of the semester, most of the writing students do is in class with pen and paper, all devices put away. To me, it’s essential that I establish baseline samples of students’ writing that I know, to the highest possible degree of certainty, is their writing. These assignments are summaries and analyses of readings, or early drafts of things that will go in their bigger, computer-drafted projects later on.

I also build in more in-class work time, where they’re working on a small assignment or their larger project — on their laptop or whatever they want — and I’m circulating around the room answering questions or just sitting down with them and picking their brains. Some students love this and others don’t, but overall it’s another way of getting familiar with their work, their thinking, their style. I also just enjoy it. One-on-one conversations are where you find students’ enthusiasms and weird passions and help them through jams. Those are the moments where students are most likely to look at me like, “Oh, that was actually helpful!”

And maybe the biggest thing I’ve changed is that I don’t teach writing online anymore. I used to teach half my load online, and now all my writing classes are face-to-face. How this has changed my schedule and what I have time for is no small thing. I’m still adjusting to it.

Do you and your students have different feelings about using AI to write?

Most definitely. But students are an incredibly varied group. Many are militantly anti-AI, and many use it daily for all kinds of things. With that said, because AI is so mainstreamed into so many environments now, their ethics around it are often different from mine. For example, many students are surprised to hear that taking something produced by AI, putting it into their own words, and turning it in as theirs could be considered plagiarism. A higher percentage than I am comfortable with would still consider that “their writing,” which of course indicates a wide gap between what teachers and students consider “writing” to be. I’ve heard teachers say that what they think of as “typing,” many students see as “writing.”

Are you more hopeful or concerned about the future of writers/writing when you teach today (in light of AI)? 

This is another question that’s complicated by how varied students are in their opinions of AI. But I think I am concerned about the future of writing done by “professionals who write,” and hopeful about writing done by “professional writers.” Most people who write in their work are not professional writers: they’re engineers, doctors, salespeople, or whatever. The writing they read and produce on the job is very much a means to an end, so why wouldn’t much of it be farmed out to AI? That’s an honest question: What forces would stop this type of writing from being largely or entirely AI produced? Personal pride? Corporate rules and regulations? Mostly, I think pinched and uncritical notions of “efficiency” will filter lots of this kind of writing through AI.

But I am hopeful that people going into fields where the writing is the point — professional writers — will largely have enough pride to retain control over their own voice. Obviously there will be pressures going the other way, like there already seems to be in journalism. But I still think commitment to the human will remain strong among this cohort of professionals.

Latria Graham

Assistant Professor, Augusta University
Augusta, GA
Writes: Scratch! 
Courses Taught: Intro to Creative Writing, The Art of the Essay, Composition 2, Magazine Feature Writing

Do you have an AI policy in your creative writing or composition courses? Is it anything other than “No AI allowed”?

The rule is no AI of any kind, and I have a one-page explanation on my syllabus about why: the environmental costs, the racial bias baked in because of who built the tech, the theft, the lack of citation, etc. I have this very firm line. I tell them I don’t even want them using AI for summaries or idea generation, because I want to see the associative way they think, and I design my assignments this way. Students also must sign a contract saying they’ve read the syllabus and understand the rules for the course. It’s part of their first graded assignment.

I do not use AI detection software in my class — I don’t want to feed their creativity to a machine. It isn’t my job to surveil, or trick them and I refuse to take that role on. The act of writing, of being in literary community, runs on trust. There are social contracts between writer and reader, writer and editor, and teacher and student. My classes are the first places students, as writers and artists, can see if they can fulfill their part of the social contract.

How much have you shifted methodology in terms of teaching writing now that AI LLMs are ubiquitous?

I teach the composition course as a food writing course and we "sample" writings. The textbook cannot be found online, only print. I didn’t do that on purpose, it’s just that was the best textbook for the work I wanted to do.

When it comes to my upper-level creative writing courses, that work engages with “the now,” and I often focus on books and art being published in the year that I’m teaching. I bring in stories that were published the week before and we tug on the meaning of that poem, article, or essay. I pair those readings with older craft essays and lauded texts that inform the lens that we’re looking at current work through. LLMs do a terrible job of understanding writers like Toni Morrison and Francine Prose. I’m in creative writing, so I don’t want literary analysis or summaries; I want interesting work pulled from the life experiences and research interests of my students that Morrison and Prose nudged them to think about. We do that critical work together as a group in a classroom.

As a working artist I’ve always talked to my students about current issues in the literary world. Lately, the current events section of the class has been about the ways that this technology is making life challenging for them. We’ve talked about what happens when literary magazines shut submissions because bots are packing their submission portals with more slop than they can read, which takes away opportunities for them and other new-to-the-game writers. This week we’ll talk about all the issues around AO3 going offline, including the fact that someone has uploaded many of those stories to another platform and is now selling the pirated work for money.

Do you think you and your students have different feelings about using AI to write? 

I think my students are trying to figure out how to give their inner worlds shape when they've had technology shoved into their faces since birth. That fact means they've also watched technology take from them, embarrass them, and document them. 

Recently, I gave them the option to turn in an assignment in their handwriting (I tell them they can scan or take photos of their work and upload it to D2L, the university's learning management system [LMS]) or the standard Microsoft Word doc. About 30% handwrote it.

Some of them tell me writing by hand is hard and tiring on their hand muscles, and that makes them think hard about what they want to say. It forces them to distill the image or idea.

But because instructors are beholden to the LMS, I think there's a lack of imagination about how to bring the human element to this conundrum. Everything is about how to make things convenient and frictionless. Personally, I print all of their assignments and then grade them in blue ink. I upload those comments into the LMS, so it’s double the work for me, all under the guise of “ease.” All of my creative writing classes would be offline if they were allowed to be.

Art can be uncomfortable. My students don't have to do their creative work the way I do mine, but they should know how artists go about the act of creation. I let my students watch me struggle. I let them watch me work out complex ideas. I am often doing my writing prompts alongside them, writing longhand on legal paper. I show them things that weren't modeled for them, because most of their instructors are not working artists. Seeing me work gives them something to pattern on. 

I think students are bored with AI because they recognize it's always being used to sell them something — they recognize that it will cost them something too. There’s also a class component to this: many of my students are working class and pinching pennies. They’re thinking about the price of services in a way that students from more affluent backgrounds might not be. 

Are you more hopeful or concerned about the future of writers/writing when you teach today (in light of AI)? 

I’m hopeful. I think the larger systems-level stuff is challenging because in order to live we need jobs, and my job is being transformed using metrics that favor technology instead of critical thinking. Some of the measurements that will determine whether or not I will get tenure hinge on student reviews which are about whether or not they “like” the professor, not whether the professor forced them to grow as a thinker and/or artist. It would be nice to have a way to resist that pressure. I think most of us working right now feel it, regardless of the role. I recognize I might lose my job because of the way I teach; I am also not changing the way I teach.

Emily Strasser

Tufts University, Professor of the Practice in Creative Nonfiction and Journalism
Boston, MA
Writes: Half Life of a Secret (University Press of Kentucky, 2024)
Courses Taught: Writing The Climate Crisis, Weirding the Essay

Do you have an AI policy in your creative writing or composition courses? Is it anything other than “No AI allowed”?

Yes, I do. It’s “don't use it,” basically, for anything other than transcribing interviews. I do tell them that I'm not interested in policing whether or not they're using ChatGPT as a search engine, but I emphasize that you have to cite everything with a real source, and I will spot check them. You're responsible if you cite something that is not reputable or doesn't say the thing you say it says.

My students are mostly pretty educated about AI. The students at Tufts who are signing up for a creative writing course, it’s not required for anything, so they want to be there. In my syllabus, I say that [LLMs are] built on stolen labor, and I'm not interested in seeing what a machine comes up with. I’m interested in seeing what you come up with and your own voice, and you're not going to develop a sense of your own style or thinking if you're really in any way letting AI influence you. 

Sometimes they're surprised by the fact that I don't want them to even use it to brainstorm, because it's suggesting certain things, and shaping the way they’re thinking through something that is not based on what they, as a human, are exposed to. And why would one, especially at this sort of formative stage in their writing, let a machine access and help shape the way that they think?

How much have you shifted methodology in terms of teaching writing now that AI LLMs are ubiquitous?

I'm going more analog. I've gone back to course packets as much as possible. It doesn't eliminate the possibility that they're going to load a reading into AI, but it's one level of friction. Instead of reading responses on Canvas, I’m experimenting with having them do a handwritten annotation on their course packets that I collect periodically, so I think that's valuable. At some point I started to let laptops in the classroom more, and now we're going back the other way. 

It's a process. There are still things I'm using Canvas for. My students do type and submit their work online. Some of my colleagues in English are going back to blue books or oral exams. But I need quiet time alone — a lot of time — to do my own writing, so I'm not going to force them into a process that I would hate and wouldn't work for me. 

It’s not realistic, if you want to try to write in the world, that you do everything by hand. For the most part, that's been okay. I'm reading their work. I'm talking to them about it. I have a sense of their voice, and when their writing is really, really good, I can tell that the students have done it, because I've engaged with them through the whole process. I think it's reading responses and workshop letters that I'm worried about more, in terms of AI use, than creative work. I have tried to add some safeguards, but they're probably not completely foolproof.

Do you and your students have different feelings about using AI to write?

I actually feel pretty lucky that most of my students are very excited about creative work and take themselves and each other seriously creatively. And when we talk about AI elements in creative writing, they are pretty horrified by it. Even the ones who use it in other classes, or for other uses, are like, “I don't want it to touch my creative work.” 

We read Vauhini Vara’s viral essay "Ghosts" from The Believer [in 2021]. It was an early ChatGPT co-written story, that is transparent about which parts are written by AI, and it’s an interesting commentary on what the machine is doing, and then this piece she wrote in Wired [in 2023] about that process, and what she learned from it. That ends up being a pretty lively discussion, and students get pretty heated. They find it freaky and uncanny and disturbing. 

I tend to feel kind of heartened, for the most part, when we talk about AI in my classes, because they're a self-selecting bunch. But I do wonder if [that statistic that] “90% of undergraduates are using AI”  is true. What does that really mean? Because it's not 90% of my students. And there were all those grads last year booing the tech execs at graduation. They’re not really happy about it. I think that [students] want chances to learn without AI, and they understand that it's doing something to their brains.

Are you more hopeful or concerned about the future of writers/writing when you teach today (in light of AI)?

That's a great question. I think being in the classroom does make me feel hopeful, and I'm even more touched when I see a student come in who's really doing the work of revision. I'm such a big revision proselytizer, and when they're really like, “Wow, I think I just gotta totally rewrite this,” or “I'm just gonna play around until I figure out what it is,” I'm like, “Yes, that's the spirit!” The fact that they're choosing it, I do find heartening. 

What I get worried about is what this does to the market. What does this do to our ability to get published and find readers? I think the desire to write is still there, and the desire to read things written by humans, but we need the money to publish them and the resources. 

Huan Hsu

Lecturer, Amsterdam University College
Amsterdam, Netherlands
Writes: The Porcelain Thief: Searching the Middle Kingdom for Buried China (Penguin, 2015)
Courses Taught: Intro to Journalism, Cre[AI]tive Writing: Literary Invention in the Age of Artificial Intelligence

Do you have an AI policy in your creative writing or composition courses? Is it anything other than “No AI allowed”?

In both the creative writing and journalism courses that I teach, I have a general policy that students shouldn't use generative AI unless explicitly instructed to do so. AI is a major preoccupation for the university, and its stance on AI use by students has shifted in the past couple of years from "no, unless" to "yes, unless." But lecturers are allowed to specify their own policies. 

How much have you shifted methodology in terms of teaching writing now that AI LLMs are ubiquitous?

Methodology not so much. My pedagogical approach to teaching has more or less remained the same. The content has changed, in that we explicitly engage with AI in my creative writing course in readings, (guest) lectures, class discussions, and in-class writing activities. I even developed a course about AI and creative writing.

Do you and your students have different feelings about using AI to write? 

The vast majority of the students in my creative writing courses aren't interested in using AI to write creatively or journalistically, which is probably to be expected since most are taking the courses outside of their major requirements and are (usually) intrinsically motivated to do the work themselves. I think even when under time pressure, they aren't turning to AI. Plus, AI and the pedagogical discussion around it has changed a lot in the past few years, and at least in my classes I have noticed an increasing antipathy towards AI. 

Until this year, we could not require students to use AI, so when we did some in-class exercises with it, I always had to offer an alternative exercise, and something like one quarter of the class of 25 would opt for the alternative — usually objecting to the environmental cost of using AI. Now that AI and the discussion around it is so pervasive, I'm noticing my students are seeking out opportunities to just get away from it. The attitudes shared in class discussions about AI and creative writing have also seemed to harden. 

This year, I didn't have a single student in a class of 30 say they thought AI had a place in creative writing; for them, creative writing was inseparable from the human. Even students in the hard/computational sciences, including students studying AI, said they didn't see any place for AI in creative writing, not least because they were aware of so many negative consequences of AI in general that they just couldn't imagine it having any positive association with creative writing. 

Maybe not relevant to the question, but my students tend to apply their biases against AI to using AI for journalism, as well, while I feel slightly differently. 

Are you more hopeful or concerned about the future of writers/writing when you teach today (in light of AI)? 

I may be unjustifiably optimistic, but I'm fairly hopeful about the future of writers and writing. What I'm a little more concerned about is the future of reading, which of course is an essential part of writing. (This was my big takeaway from that big Atlantic piece from the summer.) Sometimes struggling through a dense or difficult text is the whole point, and AI can sidestep this important process by quickly extracting whatever main points its user might think they're supposed to be looking for. 

Juan Carlos Reyes

Associate Professor of English, Seattle University
Seattle, WA
Writes: Three Alarm Fire: Stories (Hinton Publishing, 2024)
Courses Taught: The Hypoxic Fiction Workshop, Digital Fiction, Multimedia Scriptwriting, Podcasting: the Sound of Story

Do you have an AI policy in your creative writing or composition courses? Is it anything other than “No AI allowed”?

I’ve started prefacing all my courses with a conversation about large language models and AI to arrive at a collective decision about the impact we want it to have in the classroom. I start there because these emergent technologies are already so embedded in our lives that it’s no longer [about] how can we avoid them, but about gauging the impact on aspects of ourselves and our everyday [lives]. 

I’ve had well-intended lessons and syllabi hotly contested and debated in the classroom, and I’m thankful for that. Holding space for all of who we are and all of what we’re feeling with these emergent technologies is maybe the most important contribution to a whole and holistic collective conversation about this stuff.

Any understanding of creative or scholarly activity that doesn’t center an embodied experience of reality is, at least partially, modeled under capitalist notions of fear and scarcity: you have to produce and produce and produce and faster and faster, this or that is the acceptable sound or model for your work, and your production must be purposeful and bring value. That’s a fear-driven imaginative process, and taking that off the table is, I hope, also taking away the anxiety that can make us want to rely on emergent technologies and believe we have to be perfect.

How much have you shifted methodology in terms of teaching writing now that AI LLMs are ubiquitous?

I’ve shifted completely. While I’ve always understood my creative work and my imagination to be a reflection of my embodied experience in the world, I now over-emphasize as much in the classroom.

What it means to exist is being actively challenged by these advanced large language models and AI broadly. A whole slew of advanced algorithms (and their hype men and marketing consultants) keep telling us we’re just predictive text generators ourselves, which we are not.

I try to embody my way through all this in my class, even when all it looks like is the instruction “Take your notebook and pen out for a writing prompt.” This era of AI has compelled me to try more insistently to remind my students, as much as to remind myself, that we’re so much more than linear algorithms. 

Do you think you and your students have different feelings about using AI to write? 

Not my creative writing students. They have expressed time and again their distrust and their distaste for these systems, even while acknowledging that these systems comprise parts of their lives online through apps and search engines. They share how much they didn’t ask for these things and how much they don’t need them. And I think we’re all coming to terms with the pseudo-religious zeal, as we have with the capitalist ends, of how quickly and intensely these things have rolled out.

Where I believe my students and I differ is that I do believe there’s a place for experimentation with these systems. However, don’t mistake me: I firmly believe any outsourcing of the imaginative process reduces our imaginative capacity, and so much, if not every, form of expression is an imaginative endeavor.

And yet, yes, my Venn diagram is a bit wider than that: I do feel as artists and writers that creative experimentation with these algorithms in the most playful, unserious, and daring ways possible is essential. So many writers, in particular, have done some really cool work with AI agents and large language models, revealing truths about the foundations, architectures, and tendencies of these algorithms. This kind of work essentially creates social critique of the state of these systems and their relationship to humanity. 

Are you more hopeful or concerned about the future of writers/writing when you teach today (in light of AI)?

I’m as hopeful and concerned as I was when I first embarked on a messy creative life all those years ago. 

This post was edited by Rahawa Haile.

Post-Credits Scene From the Garden

Scratch is a weekly newsletter about how your favorite writers are surviving (or not). We are a four-person collective publishing essays, interviews with other writers, and topical inquiries into the publishing and media industries today. If you like what we’re doing here and want to support us, a paid subscription is the best way to do so.