Godot maintainers overwhelmed by AI-generated pull requests

Godot maintainers struggle with slop

Godot engine maintainers are warning that low-quality, AI-generated pull requests are overwhelming the project and burning out volunteers. Lead maintainer Rémi Verschelde says these "AI slop" contributions are increasingly draining and demoralizing for the team. The influx of automated code submissions is forcing project leaders to choose between being welcoming to new developers and protecting their own mental health.

Game designer Adriaan de Jongh recently sparked a wider discussion by calling these Large Language Model (LLM) submissions a massive time waster for reviewers. He noted that the changes often make no sense and come with extremely verbose descriptions that the submitters clearly did not write themselves. In many cases, the users submitting the code do not actually understand the changes they are proposing to the engine.

The Godot project relies on a community-driven model where any engine user can theoretically make an impact on the source code. Verschelde is now calling for more funding to pay additional maintainers specifically to deal with the onslaught of useless pull requests. He expressed doubt about how long the current team can sustain the manual labor required to filter out automated junk.

Open source projects create bans

The Blender 3D design project is facing similar issues and recently proposed a formal AI contributions policy to stem the tide. Blender joins a growing list of major open-source organizations that are tightening their rules around machine-generated code. These projects are struggling to maintain code quality as the barrier to entry for making a contribution drops to a single button click.

Several high-profile organizations have already implemented or are currently debating AI contribution policies, including:

• The Linux Foundation
• Fedora
• Firefox
• Ghostty
• Servo
• LLVM

The Gentoo Linux distribution is taking even more drastic steps by migrating its operations from GitHub to Codeberg. Gentoo developers cited GitHub’s continuous efforts to force Copilot usage on repositories as a primary reason for the move. They argue that GitHub’s corporate strategy actively incentivizes the behavior that is currently breaking the open-source workflow.

Developers build automated filters

Some developers are fighting back by creating their own automated tools to detect and block AI-generated content. The developer of the self-hosting toolkit Coolify recently released an Anti Slop GitHub Action to protect his repository. He claims this single tool could have closed 98 percent of the low-quality pull requests his project received over the last few months.

The Coolify developer clarified that he does not oppose AI as a technology, but rather its irresponsible use in professional environments. He stated that AI remains one of the best tools ever released when used by experienced developers who follow project guidelines. The problem arises when inexperienced users use LLMs to "spray and pray" contributions across hundreds of repositories at once.

Other maintainers are calling for GitHub to take responsibility for the environment it has created. Critics argue the platform currently rewards users for the quantity of their contributions rather than the quality. This gamification of the "green square" contribution graph encourages users to submit as many pull requests as possible, regardless of whether the code actually works.

GitHub promises new moderation tools

GitHub director of open source programs Ashley Wolf acknowledged the problem of low-quality contributions arriving at scale last week. While she avoided blaming AI directly, she admitted that maintainers are facing an unprecedented level of "noisy inbound" activity. GitHub is now working on features to make it easier for maintainers to delete problematic pull requests directly from the user interface.

Wolf highlighted several existing features that maintainers can use to protect their projects right now. These include the ability to limit pull requests to existing collaborators or to disable them entirely during high-traffic periods. Maintainers also have the power to enforce temporary interaction limits for specific users who repeatedly submit low-quality work.

GitHub is considering more advanced gating mechanisms to help maintainers manage the flow of code. These potential updates include:

• Criteria-based gating that requires a pull request to be linked to an existing, approved issue.
• Automated triage systems that use AI to identify and flag other AI-generated "slop."
• Mandatory contribution rules that must be checked off before a user can hit the submit button.
• Enhanced reporting tools for project leads to flag accounts that prioritize volume over accuracy.

The conflict of interest

There is a visible tension between GitHub’s business goals and the needs of the open-source community. GitHub’s parent company, Microsoft, has invested billions of dollars into AI and expects Copilot to be a central part of the developer experience. However, the data suggests that the widespread availability of these tools is directly increasing the workload for volunteer maintainers.

A GitHub product manager recently stated that the company does not believe counting AI-generated pull requests is the right metric for success. This stance has frustrated maintainers who see the raw number of incoming requests as their primary source of stress. The company is currently hosting official discussions on the subject to find a middle ground between promoting AI and protecting project health.

The "anyone can contribute" ethos of open source is now under direct threat from automation. If maintainers are forced to close their repositories to outsiders to stay sane, the entire ecosystem could become more insular and less innovative. For projects like Godot, the challenge is finding a way to keep the door open for humans while slamming it shut on the machines.