Generative AI in University Programming Courses

A Survey of Student Practices and Perspectives

Abstract

This study explored the optimal design and execution of programming courses in the era of generative AI. This study was based on a survey of students enrolled in this course. Specifically, this study investigated students’ actual use of ChatGPT and their perceptions of the value of learning programming. The investigation targeted students enrolled in an applied course in the Mathematical and Data Science Minor at a comprehensive university. Subsequently, the findings were used to study the implications of the designs of examinations and assignments. The survey revealed that 80% of the respondents had experience using ChatGPT, and the purposes of use varied depending on their prior programming experience. Additionally, students were apprehensive regarding the fairness of evaluating assignments and examinations when generative AI tools are employed. In response, a series of countermeasures has been proposed, including the implementation of in-person assessments that prioritize coding aptitude and reasoning and comprehension skills. Conversely, another counterargument posits that programming education should assess the ability to use AI effectively as a practical tool.