Linux Foundation Report Finds AI Cuts Coding Time by 52 Percent While Low-Quality Contributions Rise
October 27, 2026 · 3 min read
Also published in Português (Brasil)
The Linux Foundation released a new study on October 20 revealing that generative artificial intelligence has reduced the average time required for code writing tasks by 52 percent, yet simultaneously increased the burden on open source maintainers due to a surge in low-quality contributions. The report indicates that approximately 25 percent of pull requests received by open source projects are now estimated to be low-quality content often referred to as AI slop, with this figure rising to 31 percent in large-scale projects.
The research highlights a significant divergence between individual developer efficiency and community-wide workflow health. While 59 percent of maintainers reported an increase in the volume of code contributions since the rise of generative AI, this number reaches 69 percent within large projects. Maintainers estimate that one quarter of these incoming requests lack sufficient quality or validation, requiring substantial effort to assess their value. This phenomenon creates a bottleneck where the ease of generating code outpaces the capacity of volunteers to review, verify, and integrate changes safely.
Efficiency gains from artificial intelligence tools extend beyond mere code generation but do not apply uniformly across all development activities. Survey respondents reported time savings in six categories of open source development, with code writing seeing the largest reduction at 52 percent. Even communication-heavy tasks such as proposal discussions saw time savings of about 19 percent. However, for tasks involving high complexity, 21 percent of respondents stated that using generative AI actually required more time than traditional methods, suggesting that efficiency advantages in routine tasks do not automatically translate to complex software engineering scenarios.
Developers are reallocating the time saved through AI assistance rather than solely increasing code output. According to the survey, 66 percent of respondents invest their efficiency gains into managing technical debt, while 63 percent focus on developing new features and 58 percent dedicate time to learning and skill enhancement. This shift indicates that while personal productivity has improved, the overall collaborative process faces new challenges as the nature of work evolves from creation to verification and governance.
Trust and mentorship dynamics within open source communities are also shifting under the influence of these tools. Although 71 percent of maintainers agree that generative AI lowers the barrier to entry for participating in projects, 53 percent disagree with the notion that it improves community trust. The ability to generate code quickly does not ensure that contributors understand the modifications or can maintain them long-term. Furthermore, 46 percent of maintainers believe there is an increased need to mentor new contributors, while only 32 percent feel this need has decreased, pointing to a transition where guidance focuses more on code validation and quality management than initial syntax instruction.
Governance mechanisms within open source projects have not yet fully adapted to the prevalence of AI-assisted contributions. The study found that only 35 percent of projects have established formal policies explicitly allowing the use of generative AI in code contributions. Meanwhile, 32 percent of projects have neither such policies nor plans to create them. On the contributor side, 61 percent claim to perform detailed reviews and full functional testing before submitting AI-generated code, yet there remains a lack of consistent standards for disclosing AI usage. This information asymmetry forces maintainers to bear higher review costs without clear visibility into how much of a submission was generated by machines or validated by humans.
The report was published by LF Research with sponsorship from Red Hat and support from Futurewei, OpenSSF, and LF AI and Data. It concludes that while artificial intelligence accelerates code production, the associated costs of verification, security review, and long-term maintenance have not decreased proportionally. The foundation recommends that open source projects establish clear policies regarding AI-assisted contributions, including requirements for testing, documentation, and disclosure, to ensure that the speed of code generation does not compromise the reliability of software systems.