AI Code Review Cuts Inspection Time by 2.5-4.5 Days per 1K Lines, but Quality Concerns Rise

According to research by Rohan Paul covering 207 GitHub projects and 1.02 million pull requests, integrating AI agents into code review can reduce review time by 2.5 to 4.5 days per 1,000 lines of code (KLOC). However, the study reveals a quality trade-off: 78% to 94% of PRs reviewed with AI participation contain "review smells," compared to 69% to 76% in purely human reviews. The research attributes the quality decline partly to repeated assignment of the same AI reviewer identity, reducing review diversity. The analysis also found that projects with early large-scale LLM review adoption did not achieve significant efficiency gains.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments