Exploring Student-AI Interactions in Vibe Coding

Abstract

Background and Context. Chat-based and inline-coding-based GenAI has already had substantial impact on the CS Education community. The recent introduction of “vibe coding” may further transform how students program, as it introduces a new way for students to create software projects with minimal oversight. Objectives. The purpose of this study is to understand how students in introductory programming and advanced software engineering classes interact with a vibe coding platform (Replit) to build software and how the interactions differ by programming background. Methods. Participants were asked to think-aloud while building a web application using Replit. We qualitatively coded screen recordings and associated artifacts (prompts, code changes, and error logs) to construct an interaction labeling scheme capturing how students prompted, tested, debugged, and engaged with code. Findings. For both groups, the majority of student interactions with Replit were to test common cases in the prototype or use prompts to debug. Only rarely did students analyze or manually edit code. Prompts by advanced software engineering students were much more likely to include relevant app features and codebase contexts than those by introductory programming students.