Research: AI Code Review Can Reduce Review Time, But Quality Risks Cannot Be Ignored
7x24h News
Research: AI Code Review Can Reduce Review Time, But Quality Risks Cannot Be Ignored
According to the research summary published by Rohan Paul, a large-scale study covering 207 GitHub projects and 1.02 million pull requests shows that introducing AI agent review can reduce code review time by 2.5 to 4.5 days/KLOC, but at the cost of declining review quality—in reviews involving AI, 78%~94% of PRs exhibit "review smells", higher than the 69%~76% in pure human reviews. The study points out that repeatedly assigning the same AI reviewer identity is the main reason leading to the decline in review diversity. Notably, projects that introduced LLM review extensively in the early stage did not achieve significant efficiency improvements.
Navigating Web3 tides with focused insightsTechFlow news, July 27, according to a research summary published by Rohan Paul, a large-scale study targeting 207 GitHub projects and 1.02 million pull requests showed that introducing AI agent review can shorten code review time by 2.5 to 4.5 days/KLOC, but at the cost of declining review quality—in reviews involving AI, 78%~94% of PRs exhibited "review smells", higher than the 69%~76% in purely human reviews. The study pointed out that repeatedly assigning the same AI reviewer identity is the main cause of decreased review diversity. Notably, projects that introduced LLM reviews extensively in the early stages did not achieve significant efficiency improvements.
