Pair Programming in Programming Courses in the Era of Generative AI: Students' Perspective
Mario E. S. Simaremare, Chandro Pardede, Irma N. I. Tampubolon, Putri Manurung, Daniel A. Simangunsong · 2024
Context: The emergence of Generative AI (GenAI) technology presents an opportunity to enhance students' learning experience in programming courses using pair programming. GenAI can take the navigator role in student-GenAI pairing as an alternative to traditional student-student pairing. Objective: This study explored various use cases, challenges, and learning experiences IT students faced when using the student-GenAI pairing approach. Method: We integrated GenAI into CS1 and CS2 courses in one semester split into two halves: in the first half, students worked in student-GenAI pairs, while in the second half students worked in traditional student-student pairing with GenAI as an additional reinforcement. At the end of the semester, we interviewed 12 students purposefully selected out of 103 enrollments and employed a thematic analysis approach to synthesize the qualitative data. Results: We identified five distinct GenAI use cases confirming the existing studies and matching how software practitioners utilize GenAI in the industry, indicating an alignment between education and industry practice. Furthermore, we identified six challenges. One novel challenge related to the consequence of the technology is narrowing the students' learning horizons. The students also expressed a lack of engagement and empathy in student-GenAI pairing. They preferred the traditional pairing with GenAI as additional support, providing a better learning experience. Conclusion: Integrating GenAI into programming courses can enhance the learning experience, but new challenges emerge, provoking further studies to address them.