AI coding tools flood open source with low-quality contributions, driving away maintainers

AI coding tools such as ChatGPT and GitHub Copilot are flooding open-source projects with low-quality contributions, overwhelming maintainers and threatening the health of the ecosystem, according to a Financial Times investigation and supporting academic research. The problem, termed "AI slop" by developers, has become a growing burden on the volunteer maintainers who keep the internet's infrastructure running.

The FT report, based on interviews with maintainers of projects like cURL, Svelte, and Python, found that the volume of AI-generated pull requests has surged. Daniel Stenberg, cURL's maintainer, disclosed that in January 2026 he was rejecting hundreds of low-quality contributions per week, though he later noted in April that the situation had improved and AI reports were "mostly very high quality." The conflicting signals underscore the uneven impact across projects.

Research shows quality decay and contributor exodus

Separate research highlighted by TechRadar Pro documents a broader trend: online communities like Stack Overflow are experiencing a "silent knowledge reset" as expert contributors leave. The platform, a key source of training data for AI models, has seen its human participation decline sharply after the arrival of generative AI tools. A study cited in the report found that "quality decays exponentially following AI arrival," with experts increasingly unwilling to answer questions that are flooded by AI-generated responses.

• Maintainers of popular open-source projects report that up to 70 percent of new contributions are AI-generated and require significant rework or rejection.
• Stack Overflow's traffic and active answerer counts have dropped by more than 30 percent since 2023, according to data from the platform's own transparency reports.
• A Brown University economics professor found that student exam grades dropped from 96 percent to 48.6 percent when he switched from a take-home test to an in-person exam without AI access, with 18 of 86 students dropping the course.
• Researchers at the University of California, Berkeley and Chinese universities have independently replicated the finding that AI-assisted homework inflates performance while real-world skills degrade.

Long-term implications for open-source sustainability

The trend threatens the volunteer-driven model that underpins much of the internet. As Stenberg and other maintainers told the FT, the time spent triaging bad AI contributions reduces capacity for genuine improvements and community mentorship. The study's authors frame the situation as "a tragedy of the commons, where individual productivity gains externalize costs onto reviewers, maintainers, and the broader community."

Some projects are now experimenting with AI-contribution bans or stricter review processes. Meanwhile, platforms like GitHub and Stack Overflow are developing tools to detect and flag AI-generated content. The outcome may determine whether open source can absorb the productivity gains of AI coding tools without losing the human expertise that makes it viable.