AI-Assisted Coding Among Students: Measuring the Extent, Motivations, and Ethical Dimensions of ChatGPT Use in Programming Education
作者
Miguel Jusayan Puig
标识
DOI:10.1145/3802409.3802432
摘要
With the rapid adoption of Artificial Intelligence (AI) tools in higher education, platforms such as ChatGPT have become integral to programming practices among students. This study investigates how students use AI—particularly ChatGPT—during coding activities, how much of their code is AI-assisted, and how such practices influence their learning outcomes, ethics, and self-efficacy. Employing a mixed-methods design, quantitative surveys, controlled coding tasks, and qualitative interviews were combined among 150 undergraduate programming students. Results show that students frequently use ChatGPT for debugging and syntax-related tasks, with an average of 38% AI-influenced code. AI-assisted students achieved higher mean scores in debugging and problem-solving tasks but displayed lower confidence in independent coding. Thematic analysis revealed perceived benefits (efficiency, conceptual clarity) and challenges (dependency, authorship uncertainty). Findings highlight the need for responsible, transparent, and pedagogically guided AI integration in programming education.