The KDE project is currently discussing a new set of guidelines regarding the use of large language models and AI coding agents in its software development process. Project lead Nate Graham submitted the draft for community review, aiming to establish clear standards for AI-assisted contributions.
The proposed rules emphasize that AI output must be indistinguishable from human-written code. The guidelines explicitly state that contributions showing obvious or lazy use of LLMs may be ignored or closed by maintainers. The core principle is that developers must fully understand the code they submit.
The draft also introduces strict prohibitions on using AI for non-code tasks. Specifically, contributors are forbidden from using AI to generate text for merge requests, commit messages, or responses to community questions and comments. These measures aim to prevent excessive workloads for project maintainers.
KDE's approach aligns with the middle-ground strategy adopted by the Linux Kernel project, which permits AI assistance provided that human oversight remains mandatory. This stands in contrast to other open-source projects that have either fully embraced AI tools or implemented total bans on their usage.
