Understanding And Guiding Student-AI Interaction In Future Programming Education

Abstract

With recent advances in generative AI, delivering personalized learning experience in programming education has become more feasible. However, how students use AI can significantly impact their learning outcomes. To help guide student-AI interactions, it is crucial for instructors to establish effective AI usage policies. These policies may vary based on course requirements and students’ backgrounds, shaping different views on appropriate and inappropriate AI use in class. While understanding and guiding student-AI interaction is essential, it remains unclear which needs are general to programming education and which are specific to particular topics. This ambiguity makes it challenging to design a system that is both practical and useful in real-world programming courses. In this paper, we identify gaps in existing literature and propose a study to explore instructors’ perspectives on students’ AI usage. We also introduce a potential system design that allows instructors to monitor student-AI interactions, detect problematic behaviors, and intervene when these interactions conflict with the pedagogical goals.