Artificial intelligence is rapidly transforming software development, but Professor Jungpil Hahn argues that the greatest challenge is no longer how quickly code can be generated—it is how well developers can collaborate around it. As AI coding assistants become increasingly capable, they often introduce stylistic differences that are easy to overlook but can accumulate over time. Prof. Hahn’s research shows that these inconsistencies create hidden friction, making software projects more difficult to maintain and reducing the efficiency of collaboration across distributed development teams.

Drawing on an analysis of 1,817 JavaScript open-source projects and more than five million source code files, Prof. Hahn and his collaborators found that programming style inconsistency has a measurable impact on project performance. Differences in formatting, naming conventions, and code structure increase the cognitive effort required for developers to understand existing code before making contributions. This additional mental burden slows technical progress, discourages participation, and limits the effectiveness of open-source collaboration. The study also found that while modular software architecture can reduce system-wide friction, it cannot fully address inconsistencies within individual files.

From Prof. Hahn’s perspective, the findings are even more significant in the era of AI-assisted coding. As developers rely on different AI tools and prompting styles, machine-generated code can introduce diverse stylistic “fingerprints” that amplify inconsistency at unprecedented speed. He argues that coding standards, automated style enforcement, and stronger coordination practices are no longer optional best practices—they are essential governance mechanisms for AI-enabled software development. By treating coding style as a critical component of collaboration rather than a cosmetic preference, organizations and open-source communities can harness AI’s productivity while preserving the quality, maintainability, and sustainability of their software ecosystems.