AI's mixed impact on open source projects

A recent study by the Linux Foundation reveals that generative artificial intelligence (IAGen) has both positive and negative effects on open source software development. While it can save up to 50% of time on various tasks, it also introduces challenges, such as a rise in low-quality contributions.

Increase in contributions and AI-generated content

The study found that 59% of project maintainers have seen an increase in contributions since the rise of AI tools. However, 25% of the proposed changes in smaller projects and 31% in larger projects are generated by low-quality AI, often called "AI slop."

Workload and quality vary by project size

Larger open source projects face a higher average proportion of mediocre contributions compared to smaller ones. Large projects see 31% of contributions as AI-generated and of lower quality, while smaller projects have 23% and 20% respectively. Co-maintainers also face a higher rate of mediocre contributions.

Call for better project organization

Hilary Carter, senior vice president of research at the Linux Foundation, highlights the need for open source projects to adapt their structures to handle the growing number of AI-generated contributions. This includes setting clear guidelines, ensuring accountability, and providing better tools for project leads.

AI lowers barriers but increases friction

Generative AI is making it easier for people to participate in open source projects, with 71% of project leads believing it reduces entry barriers. However, it also creates challenges, with 53% of leads saying it has not improved trust within the community. Fewer than half of projects have formal policies on AI use.

Review practices and the need for mentoring

Most contributors review and test AI-generated code before submitting it, but disclosure of AI use is inconsistent. Project leads also believe that mentoring new contributors is more important than ever, with 46% saying AI increases the need for it, compared to 32% who think it decreases the need.