As mentioned at the end of this week’s class, the topic of an AI policy sprang up again. I kept thinking about the AI policies from my former colleagues that I shared during the Writing and AI class. Writing instructors, at least from my experience, approach LLMs with more resistance. Some instructors even make it a point to remove all technology in the classroom and revert to the good ol’ blue book. I tried to make a joke about blue books to my former undergraduate students, and, to my surprise, not a single student understood what I was talking about. Though, according to them, the Scantron is still alive and well! 

My colleagues’ frustration regarding LLMs isn’t exclusive to them. I met with some continuing graduate students who briefly shared their similar frustration with grading papers that lacked any students’ voices. A former colleague of mine even attempted to initiate a professional development meeting to discuss the potential for implementing AI literacy in the classroom. Most of the colleagues attending, however, did not enjoy the meeting at all. Many scrunched-up faces filled the room, and debates amongst peers lingered. 

The use of large language models in the classroom, as you can imagine, is still a highly contentious topic, especially in humanities courses. After taking this class with you all, I found myself having a bit more of an open eye regarding AI, rather than having immediate disdain that I was so used to hearing. From the student’s perspective, I find myself constantly questioning why a student immediately resorts to using generative AI for a paper. Is the workload for a particular class so overwhelming for a student that “checking off a complete assignment on a to-do list” becomes more important than relaying their own thoughts? Are students afraid or hesitant to share their personal voice, especially in a writing course? Where does that fear stem from? 

I reached out to some of my former colleagues to ask them more personal and direct questions regarding their opinions on generative AI and large language models in the classroom. I provided my questions as well as two different colleagues’ responses below.

Side note: I apologize for the lengthy blog post! I hope you all, at least, read colleague #2’s responses. I think there’s a lot of room for discussion based on their answers. I highlighted some of the more interesting statements.

Questions and Responses

1. What is your personal experience regarding AI in your classroom? What are your general thoughts regarding large language models in a writing course?
   • Follow-up: Do you think most of the colleagues in the department share your thoughts on generative AI? 
   • Should we implement AI literacy in the classroom? If so, what does that look like for you?

Colleague #1
This is a loaded question. Let me start off with my experience, which helps answer and contextualize questions 2 and 3, as well. I am opposed to any and all uses of AI in the classroom, at least for English 101. I am staunchly against AI for a few reasons. The first is its environmental impact. I feel a moral responsibility to convey the reality of its water use to students. If they still choose to use it, that’s on them. I also think the use of AI is antithetical to the humanities as a discipline. The reason I found myself teaching was because I felt passionate about getting students to critically think about the texts written by OTHER HUMAN LIVES, to cultivate empathy, and to communicate. We functionally lose that when we use AI.

I understand using AI to “polish” up writing, but I caution against that, too. I think that most 101 students are not confident in their own voice, which is why I tell them not to use it, but I don’t punish students for doing so. I understand why students whose second language (or third) language is english use that function. I think most faculty would agree with me on this. Not everyone.

Also, I can see the appeal in using AI as a research tool for 102, but AI isn’t there yet. It often makes up sources, but I think it could be a good idea to find sources that way!

Colleague #2
AI has made teaching and grading a lot more cumbersome than it needs to be.  In my experience, I suspect that most of my students are using it and finding clever ways to navigate around AI detectors or edit the output just enough that it sounds more human.  I know there are plenty of humanizers out there. 

I don’t think Large Language Models have any place in a writing course, at least not 100-level courses.  Students need to learn the basics of writing without AI.  LLMs rob them of the necessary struggle of learning because too often, students are relying on it to do the thinking for them.  This results in essays that generally all make the same arguments and claims, and students have less organic ideas.

I think most of my colleagues share the same attitudes.  I have had some conversations with senior lecturers who aren’t quite as negative about AI in the way that I am.  One senior lecturer in particular has expressed on multiple occasions that they feel we need to be teaching students how to use it ethically because they will need it in their careers.  I think this argument is weak.  It assumes that the purpose of higher education is only to prime young people for jobs, not to have them learn how to think critically and to gain a well-rounded education. It feeds into the transactional, customer service nature of higher education that I hate.

In general though, I do feel that most of my colleagues share the same thoughts and frustrations as me.

I think the term “AI Literacy” is thrown around quite a lot.  There are too many interpretations of what that actually means.  For me, if “AI Literacy” means teaching students information literacy (being able to distinguish fact from misinformation, verifying credibility, etc.) then sure, we should be teaching them “AI literacy”.  They need to know what AI is and isn’t capable of, to discourage them from relying heavily on these tools to do all the work for them. Especially since these tools are not sustainable and have many other ethical issues, which they also should be learning about. But if “AI Literacy” means teaching them how to use the tools, or teaching them prompt engineering, absolutely not.  Not in a Writing classroom.  Save prompt engineering for the STEM subjects that have a practical use for this.  

_____________________________________________________________________________________________________________

2. Why do you think students rely on large language models for a paper? How do you respond to students who say “it makes it sound better,” or something like that? 

• Follow-up: Does that explanation of “it makes it sound better” or it “fixes spelling and grammar” make you think about a student’s approach to writing? Like “hmm, why do they think this is considered “good writing?” What is good writing?

Colleague #1
Oof. I think this stems from a lack of confidence, as I mentioned earlier. But I’m not a fan of this idea of “proper English” or “good grammar” because it reeks of racist and classist undertones. I don’t know what that means when the language we use is constantly evolving. We have seen “they” and “them” develop into being grammatically correct method of referring to a singular individual. We used to get points off for that when I was in school. I believe this change is radical- and reflective of the world we live in.

I do think Grammarly can be helpful if a student (or anyone) is seriously worried about their grammar, spelling, syntax, etc. 

Colleague #2
Students rely on LLMs because everyone else is doing it, they are overwhelmed, and they are grossly underprepared for college-level writing. The literacy crisis existed long before ChatGPT became public.  I think that because we have shifted to tech-heavy classrooms and assignments, the temptation is just too strong, and the tools are too easy to access. Students haven’t been taught the things they need to know, and they lack the resilience to do things on their own because they haven’t been held accountable in the past.  I also blame grade inflation. I see a trend of young students who claim they were “straight A” students in high school, but the quality of their writing tells me that their teachers were far too generous in marking their essays, which gave them a false perception of their own mastery.  Students turn to AI because they truly don’t believe in themselves, or because they are spread so thin that they procrastinate, or they are afraid of asking for help, or they really don’t think writing matters.  It’s a lot of things. 

Students might think AI sounds better because they haven’t read enough to know what good writing actually sounds like.  This goes back to the literacy crisis I mentioned.  Many of my students in the past have told me that they don’t read outside of school, that they hate reading, or that they haven’t read a whole book in years.  A lot of work I’m seeing in my classrooms is similar to middle-school levels of reading comprehension and writing. 

They might think it’s good writing because it’s grammatically correct and has advanced vocabulary.  But to be honest, at this point I actually think most students know it doesn’t sound like good writing.  They recognize AI-generated text fairly quickly, and sometimes when we read sample essays in class I hear them say  “that sounds like ChatGPT”.  So they are more aware than not, but they still will resort to using AI at some point, even if it is only for an outline, or for paraphrasing, or for brainstorming.

They also use Grammarly because they haven’t been taught the basics of grammar.  Students can recognize when something is grammatically incorrect, but they cannot articulate why and they don’t know how to fix it, so they use Grammarly to do it for them. Too often they take the suggestions and allow the tool to completely rewrite their sentences for them. This doesn’t help their writing, and it actually changes their meaning sometimes. In general, I think most of them aren’t just using Grammarly to fix grammar.  They’re using it to generate large portions of their essays in the same way that ChatGPT and Claude and other LLMs do.

__________________________________________________________________________________________________________

3. Have you noticed a decrease, increase, or stagnant use of AI in the classroom this semester? 

• Follow-up: How can you tell a paper was written by a large language model? 

Colleague #1
I think there’s been an increase in AI this semester. It’s quite obvious. It tends to make up quotes. I also have in-class writing samples to compare.

Colleague #2
I have definitely seen an increase in the lazy and obvious use of AI.  I have about a dozen documented academic integrity violations.  I’ve never had this many in one semester. Most of these are AI related.  For context, I teach 4 courses.  These are only the ones that I was able to gather enough evidence to report. 

Large copy-and-pastes are obvious.  I can also tell when I see an overuse of the em-dash, which isn’t a smoking gun, but it raises red flags when a student never uses it in their handwritten work, but uses several of them in their typed work.  Other obvious signs are various word choices and patterns: “delve”, “underscores”, “it’s not about X, it’s Y”.  I can also tell when students are using advanced vocabulary that isn’t typical for undergraduates. The most obvious is when I ask a student to define a word they’ve used in their essay, but they can’t because they don’t know what it means/they didn’t actually read what they copied and pasted from ChatGPT.  Other signs are bullet point lists, strange emojis, shifts in font style and size, paragraphs that lack any citations or quotations, hallucinated quotations, hallucinated sources, and generally summaries disguised as analysis.

Aside from these, I can just tell. AI-generated writing lacks personality.  It’s not fun to read. It lacks depth, nuance, and creativity.  It just feels hollow. 

__________________________________________________________________________________________________________

As we develop our AI policy for next week’s class, I think it would be good to think about the concerns, comments, hesitations, and frustrations of educators who strongly oppose using of generative AI in the classroom. How do you respond to educators who completely resist the idea of implementing any sort of AI literacy in the syllabus, etc.? What would our AI policy look like that simultaneously addresses educators’ hesitation or resistance to using AI? How can we incorporate language that eases those worries? 

Thanks for reading, y’all 🙂 Good luck with your final projects, papers, and presentations!