The productivity picture is genuinely mixed. For both contributors and maintainers, generative AI is reported to lift productivity by up to 50% across all types of open source development tasks. The catch is that the same tools have made it easy to generate plausible-looking but substandard code, and maintainers say the resulting slop pull requests have pushed their workload to unreasonable levels.

The study's authors argue the findings point to a need for stronger AI governance in open source projects, including clear contribution policies, standardised attribution and verification of AI-assisted code, and continued investment in human mentorship.