Artificial intelligence tools such as ChatGPT are now common in college classrooms. At the University of Central Oklahoma, faculty and students are navigating how those tools should and should not be used.
Dr. Laura Dumin, director of UCO’s technical writing program and a member of the Regents AI committee, works with faculty navigating AI use and academic integrity concerns. Adjunct Professor Tyler Gwinn previously worked for a technologist involved in early AI development and advising Congress and the CIA, and he later invested in AI technologies through venture capital. They discussed how AI is being integrated into college classrooms and what its rapid development could mean for higher education.
The following conversation has been edited for clarity and length.
What are faculty at UCO asking you about when it comes to AI?
Dr. Dumin:
“A lot of the questions that I get have to do with, ‘I think a student cheated. How can I change my assignments, or how can I tell if the student cheated?’ So a lot of it is academic integrity issues right now.”
“For me, the big thing is that I’m giving my students guidelines, and I suggest that instructors have guidelines on all of their assignments about where AI can and can’t be used, why we want it to be used or not used in these places, how they might use it. If we’re transparent with our students, that helps students know where the guardrails are and hopefully make good choices.”
How do you decide when AI use is appropriate in a class?
Dr. Dumin:
“I think that there are a lot of different uses that are acceptable. It really depends. If the struggle is part of the purpose, if the learning is necessary for future understanding, maybe we don’t want AI in that space.”
“In some cases, the struggle to learn is what actually solidifies the learning. If you have a history paper and you have AI write the whole paper and you glance over it, you really haven’t learned much.”
Why does the acceptable-use line feel so unclear to students?
Dr. Dumin:
“I think it’s very blurry. It depends on every single assignment and what an instructor thinks is helpful use. There are times when two different instructors could disagree with each other about the exact same assignment.”
“We’ve got students who are working full-time jobs, who are full-time parents, who are trying to get work done. They’re like, ‘Maybe just this once I’m going to do this.’ I’m not saying that’s an acceptable use. I’m saying that’s the reality.”
What do students misunderstand about tools like ChatGPT?
Prof. Gwinn:
“You have to realize that it is not a database of facts. The AI the students are using, for the most part, is language models looking at recognizing language patterns. It doesn’t always give you correct, accurate information.”
“I always tell students to verify the research, because it does provide sources that do not exist or references that do not exist.”
“ChatGPT, in particular, is extremely biased. It’s biased in multiple ways — sometimes around political party, but also around countries. It has a really good bias toward Israel. Of course, its creators are Jewish, so it is interesting.”
“But again, it’s who’s utilizing it, what’s being input into it. It’s based on the data it’s trained off of.”
Where can AI genuinely help students?
Dr. Dumin:
“For neurodiverse learners, students with high anxiety, students who are first generation or not native English speakers, there’s a lot to be gained by working with an AI tool to help you understand what’s being asked of you, not to complete the assignment for you.”
“We’re not having a large language model write your whole paper. We’re saying, ‘My instructor asked me this. Can you help me understand what they’re asking? I’ve seen students take pictures of the board and say, ‘Help me organize these notes so they make sense.’ AI can be a helper tool and still allow students the space to do their own learning. It can remove some barriers.”
What concerns you most about where AI is heading?
Dr. Dumin:
“Agentic AI is a big one.”
“Students can give their D2L login, tell it to check every morning for their work, tell it to do their work, tell it to turn their work in. They’ve never touched the course.”
“If you get enough of those students using AI in problematic ways, that ends up devaluing everybody’s degree. If anybody can get a B in chemistry or engineering or English or history, then what do those degrees matter anymore?”
What will still matter in a world where AI can generate so much?
Dr. Dumin:
“If you don’t know how to talk to people, if you don’t know how to look at your audience and judge what they might want and need, if you don’t know how to have a conversation over coffee or dinner, those are the kinds of things that are not going to go away and become even more important.”
“All the things that gen eds teach — those are the things people are going to need.”
