The flood of AI-generated code is forcing open-source projects to abandon their founding principles, with developers implementing new restrictions to combat what one creator calls "AI slop." What began as a promise to democratize software development has instead overwhelmed maintainers with low-quality submissions, prompting technical countermeasures that close the open door that defined the movement.
Earlier this month, developer Mitchell Hashimoto launched a system restricting GitHub contributions to "vouched" users only. His approach effectively ends the trust-by-default policy that characterized open-source software for decades.
"AI eliminated the natural barrier to entry that let OSS projects trust by default," Hashimoto explained in his announcement.
The cURL data transfer tool recently paused its bug bounty program after being inundated with what creator Daniel Stenberg described as "AI slop." Security researchers once invested substantial time investigating vulnerabilities before submitting reports, creating built-in friction that acted as a quality filter.
"There was a built-in friction, but now there's no effort at all in doing this," Stenberg said at a recent conference.
"The floodgates are open."
Across major open-source projects, maintainers report noticeable declines in submission quality despite increased volume. Jean-Baptiste Kempf, president of the VideoLAN Organization overseeing VLC media player, observed that merge requests from developers unfamiliar with the codebase often prove problematic.
While AI tools help experienced developers work more efficiently, they enable inexperienced contributors to generate code that appears functional but lacks integration quality.
The Blender Foundation has encountered similar challenges with AI-assisted contributions. CEO Francesco Siddi noted such submissions frequently waste reviewer time and affect team motivation. The foundation hasn't banned AI tools but doesn't recommend them for contributors or core developers.
A January paper from Central European University researchers found that vibe coding, driven by AI assistants like Claude Code, Cursor and Lovable, decreases how deeply developers engage with code, documentation, libraries and other developers. Traditionally, developers select packages, read documentation and interact with maintainers and other users.
With vibe coding, an AI agent can select, compose and modify packages end-to-end while human developers may not know which upstream components were used.
The velocity at which code can now be developed threatens to flood the market with vibe-coded open-source projects, making it difficult for enterprises to discern which are well written, well maintained and backed by competent teams. Julian Gericke, chief technology officer at LSD Open, predicts this abundance will create stronger incentives for enterprises to invest in reliable, mature projects.
Open source investor Konstantin Vinogradov identifies a critical imbalance accelerating with AI adoption. Software codebases grow exponentially with increasing interdependencies while the number of active maintainers grows slowly at best.
"AI does not increase the number of active, skilled maintainers," Vinogradov remarked. "It empowers the good ones, but all the fundamental problems just remain."
Calwyn Baldwin, automation team lead at Johannesburg-based enterprise open-source solutions provider Obsidian Systems, argues that like any tool, AI coding assistants reflect the skill and care of their users. Developers with good habits get good results while those with bad habits inevitably produce bad code.
The most impact may be that junior programmers won't develop strong architectural skills because AI tools remove cognitive load associated with traditional development.
Bennie Kahler-Venter, a senior automation engineer at Obsidian Systems who distinguishes between basic vibe coding and higher-level architectural approaches like spec-driven development, says guardrails make good-quality software repeatable when properly implemented.
If engineering involves merely producing working software, then AI represents tremendous progress according to industry analysis published earlier this month on TechCrunch's platform covering these developments extensively throughout February 2026.
