Study: AI-Generated Code Contributions Reduce First-Time Developer Merge Rates 18%
Key Takeaways
- ▸First-time open source contributors saw PR merge rates drop 18.18% in 2025, despite overall increases in contribution volume
- ▸Low-quality AI-generated code ('AI slop') creates a denial-of-service effect on OSS communities by overwhelming reviewer volunteer capacity
- ▸Study identifies 11 remediation strategies validated by 229 OSS practitioners, organized into preservative, adaptive, and transformative approaches
Summary
A comprehensive research analysis of 294 open source repositories and over 2 million pull requests reveals that low-quality AI-generated contributions are significantly impacting open source sustainability. The study, titled 'AI Slop is DDoSing Open Source,' identifies what researchers call an 'AI-DDoS' effect: an overflow of plausible but low-quality AI-generated code that depletes volunteer reviewer capacity in OSS communities.
While overall PR volume increased in 2025, merge rates declined significantly, with first-time contributors experiencing an 18.18% drop in merge success rates relative to counterfactual projections. The research frames this as a sustainability crisis where communities resort to defensive review strategies that paradoxically reduce openness and create barriers for genuine newcomers.
Through interviews with OSS practitioners and surveys of 229 experienced developers, the study identifies 11 remediation strategies organized into preservative, adaptive, and transformative orientations. These strategies offer a roadmap for open source communities navigating the challenges posed by AI-generated contributions.
- Communities using short-term defensive review tactics risk creating a 'sustainability trap' that undermines long-term openness
Editorial Opinion
This research exposes a critical tension in the AI era: generative AI's ability to produce plausible code has outpaced human capacity to review it thoughtfully. The 18% drop in first-time contributor success is particularly alarming because it suggests AI tools may be raising barriers for newcomers—traditionally the lifeblood of open source communities. The finding that defensive strategies corrode openness points to a deeper concern: without intentional design and support, AI-accelerated productivity may decrease overall community health. AI companies and tool providers should view this research as both warning and call-to-action: supporting open source communities through remediation strategies is essential for sustainable AI development.
