With the popularity of AI-based tools, the landscape of learning and teaching software engineering has shifted to a new era, which has left both educators and students confused regarding the extent to which these tools are reliable, secure, and, most importantly, result in efficient student competence development. In this study, we explored how the use of AI tools such as ChatGPT and GitHub Copilot affect the performance of 36 students in nine teams in a software engineering project course. We also explore the perceptions of the students regarding the use of AI tools in software engineering. We divided the project teams into three groups based on their use of AI tools: group 1 used AI tools freely, group 2 used AI tools in a restricted manner, and group 3 did not use any AI tools. The results indicated that while all groups successfully finished their projects, AI tools were of great help in user story creation and completing a high number of features and tasks. However, groups 1 and 2 also require time to learn the AI tools and the resulting software quality was lower than that of group 3. In conclusion, AI tools like Copilot and ChatGPT can become powerful companions to software engineering students in their educational activities.
The contribution of the second author was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2023S1A5C2A02095195).
