BYU professors learn about how students are using ChatGPT

As artificial intelligence continues to evolve, educators are trying to figure out what impacts the technology will have in the classroom.

A new study conducted by professors at Brigham Young University looks into how students are using ChatGPT to search for information. They surveyed 455 students about prompts they fed the technology.

Amanda Hughes, an associate professor of computer science at BYU, participated in the study. She said some of the most common uses were for content retrieval similar to a Google search, generating content, revising or correcting their work, and evaluating the work they’ve done on an assignment. The results revealed some students used the interactive, iterative nature of the tool to have the same sort of dialogue they would with an instructor.

Data from the study reportedly did not suggest that students were using the new tech to cheat. However, professors were curious about how educators are advising students to use AI — or in some cases deter them away from using it. “Somebody said that one of their professors said that they would come and hunt them down and fail them if found out that they were using Chat GPT,” Hughes told the Daily Herald. “But then other professors were encouraging them to use it and, and trying to come up with ways that made sense that it would help augment their learning, instead of taking over.”

Hughes thinks educators may think about looking at different ways to challenge students to use their human skills as opposed to relying on technology. “Knowing that they’re going to use ChatGPT to help them with their writing, how can they challenge or push them to do something (that) allows them to use skills that they have as a human that the computer can’t necessarily do?” she asked.

Even with ChatGPT’s human-like responses, it’s not 100% reliable. Chatbots often will “hallucinate” or make up facts about things, which seldom happens when they’re asked a question. Hughes said a number of students reported experiences with inaccuracies when it came to helping them solve a problem.

“That is kind of one of my worries with ChatGPT is that if we become over-reliant on that and we think that the answers are accurate and don’t kind of push back and question when things don’t seem quite right, we could be acting on misinformation,” Hughes said.

AI likely isn’t going away, so adapting to the new technology tools will almost be inevitable for students and educators. Ultimately, BYU researchers believe it will be up to instructors how or if they decide to integrate it to supplement student assignments.

Newsletter