Godot Engine Bans AI Code Due to Low Quality: Lessons for Open Source

What happened?

The Godot Foundation, which maintains the homonymous open source game engine, has updated its contribution policies to explicitly ban AI-generated code, pull requests sent by automated agents, and AI-generated text in human-to-human communications. The measure, announced in an official blog post and covered by Slashdot, responds to a growing problem: maintainers have been overwhelmed by low-quality contributions that require tedious review and, being machine-generated, do not help train new human maintainers.

The Foundation notes that the increase in contributions is itself positive, reflecting growing interest in Godot. However, the avalanche of AI-generated code has made pull request review 'unsustainable'. The policy now requires human contributors to be responsible for the code they submit and able to fix it if it breaks. Limited use of AI is allowed for 'minor things', provided its use is declared. Any AI-generated text in human communications will also be rejected, considered a basic lack of respect; though machine translations are accepted if the original text is human.

Why is this important?

This decision is significant for several reasons. First, because Godot is one of the most popular free game engines, used by indie studios and large companies. Its stance may influence other open source projects facing similar issues. Second, because it highlights a fundamental conflict between the automation promised by AI and the need for quality and responsibility in free software. AI can generate large volumes of code, but often lacks the contextual understanding and maintainability required by a serious project. Third, because it introduces the concept of AI 'slop' in software development, a term we will likely see more often.

Godot's decision contrasts with other platforms that have embraced AI, such as GitHub Copilot or Stack Overflow, which integrate AI assistants. However, in the world of free software, where maintenance is voluntary and trust is key, the irresponsibility of automated contributions can be lethal.

Consequences for the community and market

In the short term, Godot's policy will reduce the volume of pull requests, easing the burden on maintainers. It will also encourage human contributors to engage more deeply, as they must now understand and own the code. In the long term, it could set a standard in the open source community: other projects like Blender, LibreOffice, or the Linux kernel might adopt similar policies if the trend continues.

For developers using AI as a tool, the lesson is clear: AI should be an assistant, not a substitute. A contributor's value lies not only in the code they produce but in their ability to learn, collaborate, and take responsibility. Companies that depend on Godot must ensure their developers understand this policy if they want to contribute to the project.

What should readers know?

• Godot will no longer accept AI-generated code unless for very limited tasks and its use is declared.
• Contributors must be humans responsible for the code they submit.
• Human-to-human communications must not include AI-generated text, though machine translations are allowed.
• The Foundation will review the policy periodically, but for now adopts a conservative approach.

'AI cannot take responsibility, and we can't trust heavy users of AI to understand their code enough to fix it' — Godot Foundation

This quote summarizes the philosophy behind the decision: AI cannot take responsibility, and heavy users of AI cannot be trusted to fix their own code. In an ecosystem where maintenance is voluntary, trust and responsibility are essential currencies.