University Students' Perception and Expectations of Generative AI Tools for Software Engineering
Mounika Yabaku, Sofía Ouhbi · 2024
Adopting Generative Artificial Intelligence (AI) tools in software engineering represents a shift in how tasks like coding and idea generation are approached. This paper investigates uni-versity students' perceptions and expectations regarding the use of Generative AI tools such as ChatGPT and Copilot in software engineering. To achieve this, we conducted a questionnaire study with volunteer participants studying at Uppsala University in Sweden, resulting in a total of 127 responses. These responses were about the usage preferences, motivations for adoption, per-ceived benefits, encountered challenges, and suggested improve-ment of these tools. The findings reveal that 16 % of participants have never used a Generative AI tool, while of those who have used such tools predominantly use ChatGPT3.5. Among users of Generative AI, respondents reported benefits such as code optimization and idea generation, alongside challenges such as inaccuracies in generated content and understanding user intent. Despite these challenges, participants perceive the integration of Generative AI tools as transformative for traditional software engineering practices. The results of this paper offer insights into the practical use of AI tools and suggestions for improving their functionality, thereby influencing the future direction of software engineering education.