I'm making my students read a book

A physical book!

---

The Important Work is a space for writing instructors at all levels—high school, college, and beyond—to share reflections about teaching writing in the era of generative AI.

This week’s post is by Helen Choi, who teaches writing to engineering students at USC Viterbi School of Engineering. You can read her recent piece about her experience attending the AI and Digital Literacy Institute at the University of Kansas here. You can find Helen on LinkedIn and Substack.

If you’re interested in sharing a reflection for The Important Work, you can find all the information here. —Jane Rosenzweig

---

I teach an undergraduate writing course for engineering students that is required for graduation. When I tell people about my class, I inevitably get lots of sympathetic comments like "Now that's a hard job" and "Your course evaluations must be awful." And while the answers to both comments are yes and yes, I have always defended my students. "They are so creative and expressive," I'd say. "Engineering students are amazing writers!"

But recently, I noticed that I've stopped defending my students, because I can barely remember them. In fact, the last group of students that I truly remember—their quirks, writing, hobbies, and faces— was the cohort from the 2022 fall semester. Those were halcyon days, I tell you! We had just emerged from our COVID cocoons and returned to campus. We were so happy to be with other humans and were so loud that the physics professor in the neighboring classroom would routinely send his teaching assistant to ask us to pipe down.

Then OpenAI released ChatGPT right after Thanksgiving Break, and since then it has taken up all my headspace. I’ll admit that ChatGPT caught me off-guard. I was so busy celebrating the return to campus that I hadn’t bothered to revise my COVID-era writing course with its online everything. I just skipped to campus and implemented my array of digital tools uncritically— even when it resulted in what Jenny Lederer and Jennifer Trainor describe as a "breakdown in classroom and campus community." For example, I designed an utterly soulless peer review assignment where pairs of students sitting two feet from each other provided peer review comments on a shared Google doc in silence with zero eye contact. It made me sad watching them diligently click away. I felt even sadder when, standing inches away from my students, I added my comments to that Google doc, typing "good point!!" (the exclamation points representing my dead eyes).

Max Gruber / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/

Over the past two and a half years, I have revised this peer review assignment and other Zoom-era work to be more useful for on-campus learning environments and to help me deal with LLM-related cheating. Post-ChatGPT, I’ve had scores of difficult conversations about unauthorized LLM-usage. My approach is not punitive and positioned as “learning,” but many of these interactions have been fraught with a degree of hostility and condescension that is painful to recall. For example, one student called me “pitiful” as there was no fool-proof way to prove that his essay was LLM-generated. He was right, of course; I couldn’t prove it. Instead, I had to reduce incentives to outsource thinking and find ways to teach more effectively.

And so, I made new in-class writing assignments for drafting thesis statements, outlining, and conducting research. I also added more verbal components, such as individual conferences to discuss paper topics and low-stakes oral quizzes after paper submission. These quizzes involve me sitting one-on-one with each student in my office, pulling up the student’s paper, and discussing main ideas, organization, and other writing elements. Students who write their own papers find these oral quizzes a breeze, and I found that when I better understand the nitty-gritty of students’ thinking and writing, I can provide better feedback. I also added a unit on library research methods, as I saw that students needed more support to understand the various types of sources, where they can be found, and why they need to be… real. We sometimes write on paper and sometimes on laptops, and we sometimes work in small groups and individually.

The good news is that no one has yet complained (to my face) about having to write during writing class, and when we’ve written on paper, a few students have even expressed relief in stepping away from their laptops for even a short while. The downsides have been manageable: a few hand cramps and many requests to borrow pens.

So, this fall, I’m trying another low-tech and non-digital assignment; I'm making my students read a book. A physical book! And I have big plans. We will prepare by first reading Mortimer J. Adler's "How to Mark a Book," from 1941. We will circle and star and double underline on printed pages! And we'll have something to talk about as we learn about and from each other without layers of technology to wade through. We'll sit in a circle eyeball-to-eyeball while we agree and disagree and point to pages to prove our points.

My engineering students will likely have a lot to say about the book— Karen Hao's Empire of AI, which is an insider's look at OpenAI's corporate drama and the social impacts of the company's popular LLM products. For many of my students, a job at OpenAI is a dream come true; thus, this selection is sure to engender some piquant commentary.

This 4-week reading unit will run concurrently with my library research unit. My plan is to have students do some reading and annotating at home, and in class, each student will lead a discussion on a chapter and select excerpts for collaborative close reading. The discussion will be followed by reflective writing in paper journals. By the end of the unit, the class will have examined the meaning, context, and style of Hao’s writing, parsed the nuances of positions, and examined Hao’s use of source materials. Students will then use their annotated books, reflection journals, and library research for a major paper assignment that requires them to calculate and defend their p(doom). P(doom) is the probability that AI will destroy humanity, and I bet you can guess mine.

Of course, this could also be a complete disaster. Not counting sources, Hao’s book is over 400 pages long, so I will have to somehow account for our stubby attention spans blunted by TikTok slop. And there is also the small issue of how to get physical books into their hands, as our new campus textbook program is 100% digital, but I’m working with the bookstore on this.

Even if many things go wrong, however, I’m still going to consider reading in class a win, because at a minimum, when our books are open and laptops are closed, our growth and missteps will be truly ours and our agency and accountability clear. Those tireless, tiresome algorithms and cloying instant-answer machines will have to feed on others' data inputs for a few blips. And it will be just us in the classroom— our capricious and moody selves with more questions than answers, grappling with the ambiguity that compels learning and frees us up for remembering each other.

Wish me luck!