Given permission to use AI, most college students show surprising restraint in their final essays. Students largely rely on AI for brainstorming and research rather than having it write essays for them wholesale.
Given permission to use AI, most college students show surprising restraint in their final essays
30 Comments
**Given permission to use AI, most college students show surprising restraint in their final essays**
When given permission and guidance to use artificial intelligence tools in college writing classes, students largely rely on the software for brainstorming and research rather than having it write essays for them wholesale. These findings, published in the [*Journal of Writing Research*](https://doi.org/10.17239/jowr-2026.17.03.05), suggest that students employ computerized text generators selectively to augment their learning process. The study also revealed unexpected differences in how non-native English speakers use the technology compared to their peers.
Students most frequently used the chat function to ask for help with revision, such as making sentences shorter or altering the tone. This accounted for about a quarter of the total prompts. Another highly common use was asking the program to explain course materials, define concepts, or clarify academic readings. When researchers grouped the prompts, they noticed that students asked the software to give them advice, resources, or explanations far more often than they asked it to produce text.
The chat logs also revealed a timeline of how students engaged with the tool as their assignments progressed. Most students began their interactions by asking the artificial intelligence for help with planning and locating sources. Prompts asking the machine to produce and compose writing usually occurred in the final quarter of the chat session. This indicates that direct text generation only happened after a long conversation tackling traditional phases of the drafting process.
When the team looked at the actual submitted papers, the data showed high levels of restraint among the writers. More than half of the students who participated in the pilot program chose not to include any verbatim machine-generated text in their final drafts. Across all 50 analyzed papers, only 8.2 percent of the total submitted words were flagged in blue to indicate artificial intelligence authorship. This usage fell well below the generous half-allowance permitted by the instructors.
When students did choose to paste text directly from ChatGPT into their papers, they rarely dropped in entire block paragraphs. Only about six percent of the blue text consisted of wholesale paragraph chunks. Instead, students mostly wove small, machine-generated phrases into their own original writing. The most common rhetorical purpose for incorporating this generated text was to help with discussion, analysis, and synthesis of ideas.
https://www.jowr.org/jowr/article/view/1762
Laziness has always been an option for academia but most students don’t go that route because they care about their education and want to make a good final product. AI has its uses but it can’t write a decent paper at a college level.
If grades are curved and everyone submits an AI written paper, everybody better be happy with a C
“brainstorming” is also using the AI brain instead of their own brain…
This seems like poor methodology to answer the question. AI analysis tools are not reliable at saying what is generated vs human created.
That’s nice and all but it’s reliance on search engines in general which is making everyone less intelligent and less knowledgeable. The act of reading books and articles for information teaches you far more and far better than simply asking a computer to think for you.
Most students were never using AI to wholesale write their essays because the vast majority of students follow reasonable ethical behavior. The issue has always been primarily, the edge cases of students who do use AI to wholesale write their essays and just turn in AI output as their work. Then there is the broader discussion about how much AI assistance turns your work into the AI’s work.
Unfortunately, it does seem like neither of these issues are really easy to solve. AI can’t tell if something is AI, and humans are not reliable. It is also a lot of pressure on educators if they are thinking that what they are looking at is AI and it is really demoralizing when you are always second-guessing if it is a student’s own work.
Before the widespread popularization of AI, there was definitely a move away from rigorous exams which were criticized for evaluating students on narrow aspects of lecture materials, or penalizing students who weren’t good test-takers. However, it seems like in-person exams or essays might be the only solution towards properly and fairly evaluating students in this age.
That means 44% let it do all the work
I think there are genuinely good uses for AI with regards to education, especially for study aides. If I’m studying for a test on a specific topic, I can provide a list of what is covered, and have it generate questions to quiz me. Obviously, it is far from infallible but…. Comes down to intent
This is how I use LLMs, I ask it questions trying to leave it open to give an answer indicating that it does not, in fact, have an answer.
Bro AI sucks for anything that involves actual brainpower. I’d have it sort numbers but not proofread anything.
Edit: you’re right. It’s not very good at sorting numbers. I’ve never used AI/LLM and I never will.
It’s almost like it should be used as a tool, and utilizing it that way demonstrates the true capabilities of such technology. I have never understood why rich morons want a total replacement with AI; it lacks awareness and requires human guidance in order to perform at its peak
Sure, because college students are paying to get an education. They’re smart enough to know that they won’t get anything out of having AI do it for them.
I would be interested to see how this compares to high school students who are being forced to be there.
Unfortunately the thinking part is the part I most want the students to do. I would argue that the grammer/structure is a better use for AI, and the actual ideas and research should come from the students. If they’ve written an essay with an argument generated by someone else that isn’t research, it’s copywriting
One average college students are looking to learn things, so makes sense
I take this with a grain of salt. I personally graduated with my bachelors last year and in a final group project of mine I was the only one not blatantly using AI and I had to scrub the obviously generated sections and clean up sources it had hallucinated.
When we finally get LLMs figure out, it’ll be an efficient way to talk to the data you put into it. Right now we pour the internet into them so its basically like we figured out how to make the internet talk back. Which is very cool imo, but its not AI
I would bet that the university where this study takes place is a pretty large confounding variable that makes it hard to generalize like the title suggests
Curious what these results would look like for high school students.
College students are paying to be there, and usually _want_ to learn what they’re there for, while high school can feel like you’re trapped there, forced to do work you don’t want to do, and thus seeking whatever shortcuts you can just so you can be done with it and hang out with friends and such.
That is where it really excels. Overcoming writers block and providing references (that can be cross checked for validity) is where it can really shine. It struggles in providing full essays that don’t have massive hallucinations. Short bursts of using AI for the tedious tasks, ie providing reference that would take you hours to gather, that is where LLM’s are most successful.
Brainstorming is the most basic part of thought. People are actively making themselves dumber.
Thsts a good thing. Their brains are still on!
I make my students write all of the journals in class, and I collect them.
That means I’ve seen their actual writing before any essays are due, and it is wildly obvious when they have AI write the essay for them.
Now the “problem” is whether to report them for academic dishonesty, which I don’t usually do for multiple reasons: 1. the admins make the whole process a pain in the ass, and only do anything about it like 10% of the time, even when the students are blatantly guilty. 2. Why should I waste my time policing them when they’ve shown that they don’t really care about their own education? If they want to have a a computer write everything for them, they will be the ones who can’t write, and it may or may not matter in their lives…. I just don’t have the bandwidth to hold their hands about it.
Granted I’m from somewhere else but this is unsurprising even when I was in uni a lot of the time if people were using ai it was to brainstorm for assignments however they would write their assignments themselves but a lot us did our assignments on our own cause most of us truly did care about education and what we got tbh.
It’s also a helpful editor. I am terrible at editing my own work, and don’t always want to bother someone at 3am to go over my essay before submission. Posting my writing and the grading rubric can be extremely helpful.
if I was in college ai would’ve just been used to help me with specific topics instead of having to spend an hour searching the library for a specific book which is always put in the wrong section
Asking chat to brainstorm and revise your writing still boils down to using chat to think and write for you, whether or not you copy and paste the text verbatim makes no difference to the actual process of learning. But maybe that’s just me, a Luddite yelling at kids to get off the lawn.
These bits about the study tho:
>Madsen Hardy and her team recruited 50 of these students to share their essays. A subset of 34 participants also provided the chat logs detailing their interactions with ChatGPT during the drafting process.
“We’re concerned about AI usage, so we’re going to look at how students use AI. Any volunteers?” Selection bias much?
>To track how the software was used, instructors required students to highlight any word-for-word machine-authored text in a blue font on their submitted assignments.
“Okay why don’t you just tell us where you did no work of your own and just copy and pasted what chat said?” Self-serving bias much?
I always liked writing essays. Trying to parse knowledge with sources to make a point the teacher will understand and appreciate is a fun puzzle. I could see this.
What a horrible take. “Most” isn’t good enough.
As a teacher, this doesn’t make me any more comfortable with my students using AI or an AI-assist to write their essays.