LF Research report finds 25% of pull requests may be AI slop
Key Takeaway
A Linux Foundation report estimates that a quarter of incoming pull requests are low-quality AI-generated content, with higher rates in larger projects.
Summary
A report from LF Research, sponsored by Red Hat, surveyed 234 contributors and maintainers regarding generative AI's impact on open source. The findings suggest that roughly 25% of pull requests are low-quality AI 'slop', rising to 31% in projects with over 50 contributors. While AI speeds up code authoring, it is increasing maintainer workload and burnout concerns.
Impact of AI on maintainers
- Large projects (50+ contributors) see a 31% estimate of AI slop.
- 64% of maintainers report that code reviews now take longer.
- 63% of maintainers cite burnout as a top concern, followed by contribution quality (60%).
- 35% of projects have formal AI use policies, while 32% have none and no plans for one.
- 26% of projects lack any way to verify if a contribution was AI-assisted.
- Code authoring takes 52% less time on average, but complex tasks can take longer with AI.
Why it matters
The rise of AI-generated contributions is creating a tension between increased developer velocity and the increasing burden of quality control and review for open source maintainers.
Original reference: feed.itsfoss.com →