Use of Generative AI Tools in Software Quality Assurance Work

Abstract

This innovative practice full paper describes the authors' experiences introducing the use of generative AI (GenAI) in a hybrid undergraduate video game project course. On our campus hybrid courses have both in-person and online students participating in a single course taught by the same instructor. Online students often struggle with resolving technical issues that arise in the development of software products. The authors argue that use of GenAI in commercial software development is inevitable. We believe that students should be given opportunities to experiment with GenAI tools to complete some software engineering tasks. In this paper, the authors discuss how they integrate ChatGPT into the debugging process. The authors created a module on prompt engineering and the use of ChatGPT to help determine the causes of video game defects. The students were given two weeks to correct issues in a broken Unity 3D scene. Students used ChatGPT to generate defect repair recommendations. Students were asked to use reflective writing to assess the quality of the ChatGPT recommendations. Most students were able to correct all five game defects by entering zeroshot prompts to ChatGPT. Online students received a smaller number of suggestions from their prompts than the in-person students. Most students felt that ChatGPT helped them complete the debugging task more efficiently and enhanced their understanding of the Unity 3D environment.

Ähnliche Arbeiten