An Oracle at Our Fingertips
Regulation, Learning, and the GenAI Gap in CS Courses
Generative AI tools such as ChatGPT have found their way into students’ academic endeavors, and instructors have responded in very different ways, from complete prohibition, to moderate guidance, to mandatory usage, often within the same program. It is clear that students are using these tools, but it is not clear how students feel about their own use, and because these tools are everywhere, students are making moral and academic decisions about them every day with little guidance. Are they using AI more than they would choose to, less than they would choose to, or about as much as they would like? This study will survey undergraduates at the University of Minnesota who took a computer science course in fall 2026 and measure the gap between how much they used AI for their coursework and how much they would have preferred to use it. It will also ask what students believe is driving that gap, whether that is instructors, peers, career expectations, cost, convenience, or a feeling that they cannot do the work without it, as well as whether the amount they used felt like their own decision and whether they think it helped or hurt their learning. Drawing on Self-Determination Theory, the study will examine why students use AI in the first place, and whether that use feels like something they have chosen and endorse or something they feel pressured into. Small focus groups will then follow up with students whose survey answers show different patterns. The goal is to give instructors, departments, and students themselves a clearer picture of how students are actually experiencing AI in their education, so that the decisions we make about it are based on what students tell us instead of what we assume.
Read the full description of the study
Status, November 2026: the study is at the proposal stage.