Perceptions and Challenges of AI-Driven Code Reviews: A Qualitative Exploration of Developer Experiences

AI-driven code review tools represent transformative advances in software development, improving efficiency, productivity, and accuracy in code reviews. Despite these potential benefits, concerns about trust, reliability, and contextual comprehension persist, limiting their widespread adoption. This qualitative study explores software developers' perceptions and challenges associated with AI-driven code review tools. Through semi-structured and thematic analysis involving software developers, technical leads, and architects, the study identifies central themes, including trust in AI-generated recommendations, impacts on developer productivity, ethical considerations, and contextual awareness. While participants acknowledge the efficiency gains and educational value provided by AI tools, skepticism remains regarding the tools' ability to interpret complex business logic and domain-specific scenarios. Participants advocate for enhancements in AI-driven tools, highlighting the need for improved contextual awareness, transparency, ethical integration, and seamless workflow integration. This research adds valuable empirical insights to ongoing discussions in software engineering literature, emphasizing AI-driven code reviews as complementary tools that augment human expertise in software development processes.