CodeRabbit

AI-generated pull requests contain approximately 1.7 times more issues on average than human-written pull requests based on an analysis of 470 open-source GitHub pull requests.

Performance inefficiencies, such as excessive I/O, appear nearly 8 times more often in AI-generated code compared to human-written code.

Security vulnerabilities in AI-generated code increase by 1.5 to 2 times, particularly in areas such as improper password handling and insecure object references.

Over 90% of developers report using AI coding tools to enhance productivity and manage routine tasks.

Critical and major defects are up to 1.7 times higher in AI-authored changes compared to human-authored changes.

Code readability problems in AI-generated code increase by more than 3 times, with notable issues in naming and formatting inconsistencies.

Logic and correctness issues increase by 75% in AI-generated code, including business logic errors, misconfigurations, and unsafe control flow.