IMPACTS OF GENERATIVE ARTIFICIAL INTELLIGENCE ON SOFTWARE DEVELOPMENT: PROFESSIONAL PERCEPTIONS AND EXPERIMENTAL EVALUATION

Abstract

Generative Artificial Intelligence has been widely incorporated into the software development process, influencing activities related to programming, debugging, learning, and productivity. In this context, this study aims to analyze the impacts of using Generative Artificial Intelligence in software development, considering both the perceptions of educators and industry professionals and the results of a comparative experiment. The research is characterized as an exploratory and descriptive case study using a mixed-methods (qualitative-quantitative) approach, structured in two stages. In the first stage, an online questionnaire was administered to 11 educators and software development professionals to investigate their perceptions regarding the use of Generative Artificial Intelligence. In the second stage, a practical experiment was conducted with three Software Engineering students from the State University of Pará, in which each participant developed the same software application under two distinct scenarios: without and with the assistance of Generative Artificial Intelligence tools. The results revealed a predominantly favorable perception regarding the use of these technologies, particularly concerning increased productivity, error reduction, support for learning new technologies, and the need for professional upskilling. The experiment showed a reduction in average development time from approximately 1 hour and 15 minutes to 38 minutes, while maintaining the implementation of all planned functionalities. It is concluded that Generative Artificial Intelligence serves as a supportive tool capable of optimizing the software development process without replacing the developer's technical knowledge, thereby reinforcing its potential to support programming activities in both academic and professional contexts.