AI cuts coding time, but it's also filling open source's inbox with crap
Maintainers reckon a quarter of incoming pull requests are slop, rising to 31% on larger projects
A recent survey from the Linux Foundation reveals that while generative AI is speeding up coding for open source developers, it is also inundating maintainers with subpar contributions to deal with. A quarter of incoming pull requests are deemed low-quality AI-generated code. Open source projects have been facing a myriad of challenges related to AI-generated contributions, such as identifying vulnerabilities, dealing with false positives, and tackling duplicate reports.
Linux founder Linus Torvalds views AI tools positively, but still maintains skepticism towards low-quality submissions.
To handle the surge, the foundation advocates for the establishment of clearer policies and superior tools for projects. Nearly three-quarters (71 percent) of maintainers believe AI has lowered the barriers to open source contributions, whereas a bit over half (53 percent) disagree that the technology has bolstered community trust. The survey indicates that AI reduces code-authoring time by an average of 52 percent. However, for highly intricate tasks, 21 percent reported experiencing longer durations.
The issue of governance also surfaces, with only 35 percent of projects having a formal policy permitting AI usage. A further 32 percent have no policy at all. This lack of clear guidelines leaves community expectations somewhat ambiguous. Hilary Carter, SVP of Research at the Linux Foundation, asserts that open source projects must design to accommodate an increasing influx of AI-generated contributors, rather than merely reacting to the situation.
This involves setting clear expectations for AI usage, placing responsibility for testing and disclosure on contributors, and providing maintainers with enhanced tools for managing incoming pull requests.
Among larger projects, the percentage of pull requests identified as AI slop escalates to 31 percent. The burden extends beyond mere code changes. The curl project recently decided to make low and medium severity bugs publicly accessible to alleviate the burden on its overburdened security team, which is struggling to manage a deluge of AI-assisted reports. The project's maintainer, Daniel Stenberg, laments the strain on key maintainers, feeling they are pushed to their limits.
Carter emphasizes that getting these fundamentals right can enable AI to widen participation in open source without overloading and increasing costs for maintainers.
Written by urgent.news from The Register Software's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
