Exploring GenAI as a Tutoring Tool: A Case Study in First-Year Computer Programming
Liam Stienstra, Abdallah Mohamed, Mostafa M. Mohamed · 2025
This study explores the potential of generative artificial intelligence (GenAI) to supplement traditional teaching methods for first-year computer science students. The study evaluates GenAI effectiveness in teaching a new programming concept and examines its perceived utility from the students' perspective. A conversational AI platform, powered by ChatGPT, was developed as a pilot, and used as a virtual tutor, guiding students through an interactive, time-constrained learning session that involved two steps: teaching the new topic followed by problem-solving activities. After the session, students wrote a quiz without assistance to evaluate their newly acquired knowledge. A survey was then administered to collect student perceptions and analyze their interactions with the AI tutor. The preliminary findings suggest that GenAI tutors can effectively support learning, with most students preferring AI over human tutors due to convenience, accessibility, and stress-free environment. While variability of prompt effectiveness and students' misuse or overreliance on AI highlight areas were observed, these results highlight avenues for future research to refine GenAI strategies and improve student learning.