Two Alfred University seniors spent a semester asking a question every analytics program in the country is now wrestling with: what actually happens when students hand their coding work to AI? Their answer is honest and a little uncomfortable: the tools speed things up, but the connection to the work fades.
Jack Meyer, a data analytics and computer science major from Bloomfield, NY, and Andrew Maguire, a dual major in business analytics and finance from Allegany, NY, presented their study, "A Qualitative Study of University Student Experiences Regarding the Impact of Artificial Intelligence on Coding Coursework," at the university’s Undergraduate Research Forum held Thursday, Aug. 22.
The project began as an independent study with Yavuz Keceli, assistant professor of digital journalism, putting 30 students through coding assignments in Analytics 2, Alfred's R programming course, and finding out how much AI was involved in the process of studying. Each student completed the assignments two ways: with AI tools and without. Once the work was in, Meyer and Maguire brought the group into a focus group and listened. They wanted to know how the work feels like with AI in the loop and what changed when students had to think it through alone?
Most students said AI helped them reach a finished product faster, but the problems showed up in two places. First is trust. Students reported AI tools returning hallucinated functions, broken syntax, and answers that looked correct but didn't run. Sorting usable output from bad became its own task and was sometimes more time-consuming than writing the code from scratch. Second is ownership. Students said they felt disconnected from work they hadn't really written themselves. They turned things in and walked away feeling like they had achieved less, even when the grade landed where they wanted it.
The findings pointed to two things that mattered most. First, working with AI in coding requires a skill set closer to editing and verification than to writing. And students reported learning less when they leaned on the tools, regardless of whether the final product was correct. Meyer and Maguire’s study gives faculty something concrete to work with when redesigning assignments and gives students a clearer read on what they're trading when they outsource the thinking.
Both Meyer and Maguire graduate this semester. They spent January through early April interning in analytics and are now targeting full-time roles. The research gives them something specific to bring to a first job, with time spent stress-testing what AI can and cannot do in the kind of work they want to build a career around.
Story by Andrii Maltsev ’27
