Open Source Maintainers Say a Quarter of Pull Requests Are AI Slop
By @sharedot · · 8 pages
- Programming
- Open Source
- AI Slop
- Developer Tools
A Linux Foundation survey finds maintainers estimate 25% of incoming pull requests are low-quality AI slop, rising to 31% on large projects.
What the survey found
Generative AI is saving open source developers real time while simultaneously flooding their inboxes with low-quality contributions, according to a Linux Foundation survey covered by The Register.
Respondents estimated that about a quarter of incoming pull requests — proposed code changes — are low-quality "AI slop," and that share climbs to 31% on larger projects.
At the same time, respondents credited AI with cutting code-authoring time by 52% on average, though 21% reported that highly complex tasks actually took longer.
The survey captures a paradox: the same tools that accelerate committed contributors also make it trivial for low-effort submissions to arrive at scale.
Why this is a reversal
The Register reports that 71% of maintainers say GenAI has lowered the barriers to contributing to open source, yet just over half (53%) disagree that the technology has improved community trust. Even Linus Torvalds has warmed to AI tools while remaining openly critical of low-quality submissions.
Lower contribution friction was supposed to be unambiguously good; instead, it is diluting trust in the contribution pipeline itself.
Governance hasn't caught up
Policy is lagging badly behind practice. According to The Register's coverage of the survey, only 35% of projects have a formal policy permitting GenAI use, while 32% have no policy and no plans for one — leaving expectations unclear across most of the community.
Linux Foundation SVP of Research Hilary Carter argues projects must "design for a growing AI-generated contributor base, not simply react to it," which she says means setting clear GenAI expectations, putting responsibility for testing and disclosure on contributors, and giving maintainers better triage tools.
Her bottom line: if those fundamentals are right, AI can expand participation without overburdening maintainers.
The cost is already visible: curl
The workload extends beyond proposed code changes into security triage. The Register reports that the curl project this week announced it will move low and medium severity flaws into public view to protect a security team almost overwhelmed by AI-assisted reports.
Project maintainer Daniel Stenberg said: "We are struggling as a project as all key maintainers feel strained, overloaded and brought to the brink of what we can muster."
For a library embedded in billions of devices, the collapse of private security triage capacity is a genuine supply-chain risk, not just a maintainer morale problem.
Tooling strikes back
If AI inflates the incoming stream, other AI-era tools are trying to filter it.
Help Net Security profiles DevGuard, an OWASP Incubating open source vulnerability-management platform that wraps secret scanning, SAST, dependency CVE checks, IaC checks, container scanning, and license checks in one CLI, ingests CycloneDX SBOMs or SARIF results, and syncs findings with GitHub, GitLab, and Jira.
Its Dependency Firewall even proxies the npm, Go, and Python registries to block known-malicious packages before download. Meanwhile, Help Net Security's pack context notes how much of that noise arrives via automated scanning — the same duplication and false-positive problem maintainers cite.
What comes next
The direction of travel suggests maintainers will get formal contribution contracts, mandatory testing-and-disclosure rules for AI-assisted submissions, and better review tooling before the slop wave recedes — the Linux Foundation's own prescription, as The Register reports.
The open question the survey leaves hanging: can the community absorb a growing AI-generated contributor base without the trust signals — reviews, tests, maintainer goodwill — that make open source work degrading?