AICodeReview: Advancing code quality with AI-enhanced reviews
Yonatha Almeida,
Danyllo Albuquerque,
Emanuel Dantas Filho,
Felipe Muniz,
Katyusco de Farias Santos,
Mirko Perkusich,
Hyggo Almeida,
Angelo Perkusich
Affiliations
Yonatha Almeida
Accenture, Paraiba, Brazil
Danyllo Albuquerque
Research, Development and Innovation Centre of Federal University of Campina Grande (VIRTUS/UFCG), Paraiba, Brazil; Federal Institute of Paraiba (IFPB), Paraiba, Brazil; Corresponding author.
Emanuel Dantas Filho
Federal Institute of Paraiba (IFPB), Paraiba, Brazil
Felipe Muniz
Federal Institute of Paraiba (IFPB), Paraiba, Brazil
Katyusco de Farias Santos
Federal Institute of Paraiba (IFPB), Paraiba, Brazil
Mirko Perkusich
Research, Development and Innovation Centre of Federal University of Campina Grande (VIRTUS/UFCG), Paraiba, Brazil
Hyggo Almeida
Research, Development and Innovation Centre of Federal University of Campina Grande (VIRTUS/UFCG), Paraiba, Brazil
Angelo Perkusich
Research, Development and Innovation Centre of Federal University of Campina Grande (VIRTUS/UFCG), Paraiba, Brazil
This paper presents a research investigation into the application of Artificial Intelligence (AI) within code review processes, aiming to enhance the quality and efficiency of this critical activity. An IntelliJ IDEA plugin was developed to achieve this objective, leveraging GPT-3.5 as the foundational framework for automated code assessment. The tool comprehensively analyses code snippets to pinpoint syntax and semantic issues while proposing potential resolutions. The study showcases the tool’s architecture, configuration methods, and diverse usage scenarios, emphasizing its effectiveness in identifying logic discrepancies and syntactical errors. Finally, the findings suggest that integrating AI-based techniques is a promising approach to streamlining the time and effort invested in code reviews, fostering advancements in overall software quality.
