Use of Generative Artificial Intelligence in the Education of Software Verification and Validation

Since the introduction of generative artificial intelligence (GenAI), education in computer science has prompted efforts to incorporate it into the educational curriculum. This innovative practice full paper presents a study into using GenAI to enhance student learning of software engineering. It outlines the initiatives to introduce GenAI into a graduate-level software engineering course in Software Verification and Validation (SV&V). The paper presents the educational goals, methodologies and findings of these endeavors in this course. The primary education goal of this course is that students have a solid understanding of principles and practices of software quality assurance and seek to introduce students to diverse techniques employed for SV&V. The study presented in this paper centers on the practical application of GenAI within the domain of testing strategies. The paper introduces the findings of an exercise where GenAI was used to apply testing strategies for unit testing. The exercise consisted of the use of GenAI in the development of unit tests for an algorithm. Rigorous assessments were conducted to gauge the effectiveness of the unit tests developed for validating the accurate implementation of the algorithm. This exploration shed light on the tangible impact of GenAI on the precision and efficiency of unit testing procedures. The findings underscore the significance of encouraging students to actively explore emerging trends and methodologies in the realm of software verification and validation. By incorporating GenAI into the educational framework, students not only gain insights into the capabilities and limitations of this technology but also foster a mindset of continuous learning in software quality assurance. The paper demonstrates that it is not sufficient to use the test cases developed by GenAI for software validation since test cases recommended by GenAI do not cover corner cases which causes gaps in coverage in unit testing. The majority of the students were able to understand the limitation of GenAI in SV&V but appreciated its support in suggesting test cases for the most common cases. This exercise allowed students to enhance their creative problem-solving through human-guided AI partnership which is pivotal in cultivating a new generation of professionals capable of contributing to the ongoing evolution of software quality.