AI-Driven Acceptance Testing: first insights exploring the educational potential for test analysts

• Caio Jordan de Lima Maia UFPB
• Yuska Paola Costa Aguiar UFPB

Resumo

This article examines the potential of AI-based natural language processing models, such as ChatGPT and Bard, as educational tools in the context of acceptance test case creation. The principal findings indicate that test cases manually generated by two QA professionals (one junior and one senior) exhibited reduced coverage and accuracy in comparison to those generated by the AIs, which, while expeditious, also demonstrated limitations. The findings indicate that AI can significantly enhance the learning process for less experienced QA professionals by providing clear and structured examples. Furthermore, collaboration between AI tools and experienced professionals optimizes the process, offering greater efficiency and test coverage. The study underscores the importance of integrating technological advancements into educational practices to better prepare future software quality professionals.

Palavras-chave: Software Quality Education, Generative AI-Supported Education, Acceptance Test, Test Plan, Quality Assurance, ChatGPT, Bard