Abstract:

This study examines how students engaged with generative AI tools over the course of a technically demanding class in which most learners had little prior experience with coding or analytical software. Because formal tutoring was limited to pre-recorded videos with no opportunity for interaction or clarification, students increasingly turned to freely available or low-cost AI systems as a more responsive source of support. Using a three-wave survey administered before the midterm, during the major course project, and again in the final examination period, we track how students experimented with AI, refined their prompts, and relied on iterative trial-and-error to navigate challenging material. The findings show a clear shift in their learning behavior: initial, unstructured reliance on AI gradually developed into more deliberate and methodical use as students incorporated feedback from earlier assessments and discovered how to work more effectively with AI tools. Rather than treating AI as a shortcut, students increasingly adopted it as a practical means of understanding complex content and compensating for gaps in prerequisite knowledge. The study illustrates how course difficulty, assessment structure, and the nature of available instructional support collectively shape students’ strategies for self-directed learning with generative AI.

Keywords:

Artificial Intelligence, Generative AI tools, Motivation, Learning Tools.