The Godot Foundation and maintainers of the open-source Godot game engine were preparing stricter contribution rules that would reject AI-authored code and pull requests submitted by AI agents, according to a report by PC Gamer. The planned policy also extends beyond code to AI-generated text used in communication between contributors and project maintainers.
The change follows months of discussion about the cost of reviewing a growing number of low-effort submissions. Godot's maintainers had said in February that AI-generated pull requests were creating work that was increasingly draining and demoralising for reviewers. By the July 1 event date, the Foundation had concluded that the problem was unlikely to disappear without a formal response.
The Foundation's stated objective is not simply to reduce the number of incoming pull requests. It also wants to preserve a process through which new contributors can learn from feedback and eventually become maintainers themselves. Although a large review queue reflects growing interest in using and contributing to Godot, maintainers said AI-authored submissions change the value of the time spent reviewing them. Feedback that is consumed by a machine does not mentor a potential future project member.
Under the approach described in the report, contributions would need to come from people who can answer for their work and repair it if problems emerge. The Foundation argued that AI systems cannot assume that responsibility. It also expressed concern that people who depend heavily on generated code may not understand it well enough to address defects after submission.
The coming policy was expected to explicitly reject AI-written code while permitting assistance with limited, menial tasks. Contributors using such assistance would have to disclose it. That distinction leaves room for narrow tooling support but places authorship, comprehension and accountability with the human submitting the change.
Godot also plans to reject generated prose in human-to-human exchanges. The Foundation characterized direct, human-written communication as a matter of respect within the project. Machine translation would remain acceptable when the underlying text had been written by a person, separating language accessibility from automated authorship.
The restrictions form part of a wider attempt to make reviewing contributions sustainable. The Foundation wants to add barriers to low-effort submissions, give maintainers stronger reasons to continue reviewing code and keep the contributor pipeline useful for developing future project leaders.
The position is not described as permanent or immune to revision. The Foundation said AI tools are changing rapidly and that it would maintain a conservative policy while reconsidering the rules as the technology develops. For the immediate contribution process, however, the intended standard is clear: a human author must understand the work, disclose permitted assistance and remain available to take responsibility for the result.
