Godot, the open-source game engine powering thousands of indie titles, will no longer accept code, text, or documentation authored by AI tools. The Godot Foundation announced the policy shift after maintainers reported being overwhelmed by low-quality automated pull requests. Juan Linietsky, the project’s lead developer, stated plainly: “We can’t trust heavy users of AI to understand their code enough to fix it.”
TL;DR: The Godot Foundation now rejects code, text, documentation, and pull requests authored by AI tools, citing an inability for contributors to take responsibility for automated submissions. Maintainers stated they cannot trust heavy AI users to understand their code enough to fix it. The ban covers AI agents, generated code, and AI-generated text used in human communications between developers.
Why Did Godot Ban AI-Generated Code Contributions?
The Godot Foundation banned AI-generated code because maintainers found themselves drowning in automated submissions they described as “slop.” According to PC Gamer, the volume of low-effort, machine-generated pull requests had become unsustainable for the volunteer-driven project. Maintainers spent more time reviewing and rejecting AI code than evaluating genuine human contributions. The burden was real.
Juan Linietsky explained the core problem directly. He stated that contributors relying heavily on AI tools frequently cannot explain or fix the code they submit. When maintainers ask questions about implementation details, the submitters often have no idea how the generated logic works. This creates a support nightmare for reviewers.
The policy update addresses a growing frustration within open-source communities. Projects like Godot rely on contributors understanding what they submit. When an AI agent writes the code, the human submitting it becomes a bottleneck rather than a collaborator. Maintainers cannot have technical discussions with an automated tool. They need dialogue with someone who grasps the underlying architecture and can iterate on feedback.
The decision reflects a broader tension in open-source development. AI tools can produce functional-looking code quickly, but that code often lacks the nuance required for production systems. For a game engine used by developers worldwide, code quality and maintainability matter deeply.
What Exactly Does the New AI Policy Prohibit?
The new policy prohibits several specific categories of AI-generated content. According to DualShockers and GamingOnLinux, the ban covers AI-authored code, pull requests generated by AI agents, and AI-generated text used in developer communications. The scope is deliberately broad.
Specifically, contributors cannot submit code produced by tools like ChatGPT, Claude, Gemini, or GitHub Copilot without explicit disclosure and verification. The policy extends beyond raw code. Documentation, comments, commit messages, and PR descriptions generated by AI also fall under the restriction. The foundation wants human-authored communication throughout the entire contribution pipeline.
The following submissions are now prohibited under the updated guidelines:
- Code written by AI coding assistants including Copilot and similar tools
- Pull requests created by autonomous AI agents or coding bots
- AI-generated documentation and technical text submitted to the repository
- Commit messages authored by language models or automated systems
- PR descriptions and review comments produced by AI text generators
- Bug reports generated by AI tools without human verification and editing
- Test code and fixtures created entirely through automated generation
- Refactoring patches produced by AI without human understanding of changes
| Prohibited Content | Example | Enforcement Level |
|---|---|---|
| AI-authored source code | Copilot-generated functions | Immediate rejection |
| AI agent pull requests | Autonomous bot submissions | Immediate rejection |
| AI-generated documentation | Model-written API docs | Request rewrite |
| AI-authored communications | ChatGPT PR descriptions | Request rewrite |
Contributors who use AI tools for learning or reference are not automatically banned. The policy targets submissions where AI is the primary author and the human contributor cannot take responsibility for the output. Disclosure is mandatory if any AI assistance was used during development.
How Will Maintainers Detect AI-Authored Pull Requests?
Maintainers will detect AI-authored pull requests through a combination of manual review, community reporting, and behavioral patterns. The Godot team has not deployed automated AI-detection tools, citing their well-documented unreliability. Instead, reviewers rely on their experience identifying common signs of machine-generated code. Human judgment leads the process.
Several indicators help reviewers flag suspicious submissions. AI-generated code often exhibits specific patterns: overly verbose variable names, excessive inline comments restating obvious logic, and a uniform style that lacks the idiosyncrasies typical of human programmers. Reviewers also watch for commit histories showing suspiciously rapid generation of complex code.
The most reliable detection method is direct conversation. When maintainers question contributors about their code, AI-dependent submitters frequently struggle to explain their reasoning. They may provide vague responses or paraphrase comments without demonstrating actual understanding of the implementation. This communication breakdown is a strong signal.
According to Let’s Data Science, the policy emphasizes contributor accountability over technological enforcement. Maintainers ask whether the submitter can debug, modify, and extend the code they provided. If the answer is no, the submission gets rejected regardless of its technical merits. The foundation prioritizes human comprehension over code that merely appears correct.
Reviewers also monitor for volume anomalies. A contributor submitting dozens of complex pull requests per day raises immediate suspicion. Human developers need time to write, test, and understand code changes. AI tools can generate patches at inhuman speeds, and that velocity itself becomes a detection signal for overwhelmed maintainers.
What Does ‘AI Cannot Take Responsibility’ Mean for Open Source?
The phrase “AI cannot take responsibility” captures a fundamental challenge for open-source projects accepting AI contributions. When code causes bugs, introduces vulnerabilities, or breaks existing functionality, someone must answer for the problem. An AI tool cannot be held accountable. A contributor who does not understand their submission cannot meaningfully respond to issues either.
In traditional open-source development, the contributor owns their code. They answer questions during review, fix problems identified by maintainers, and support their contribution after merging. This accountability model breaks down when the contributor functions as a relay for AI output. They cannot provide the technical depth that maintainers need.
The Godot Foundation’s stance signals a potential shift across the open-source landscape. Projects depend on trust between contributors and maintainers. That trust requires shared understanding of the codebase and the ability to engage in substantive technical dialogue. AI-generated submissions bypass this relationship entirely.
Linietsky’s statement about heavy AI users is particularly pointed. He expressed that contributors who rely extensively on AI cannot participate meaningfully in the collaborative process that open source requires. They submit code but cannot iterate on it based on feedback. They cannot debug issues that emerge during review. They cannot explain design decisions. This absence of engagement undermines the entire contribution model.
The policy raises questions about the future of AI in open-source development. Can projects find a middle ground where AI assists contributors without replacing their understanding? Should disclosure requirements extend to all AI-assisted work? The Godot team has drawn a clear line for now. Contributors must understand what they submit.
How Does the Ban Affect Godot’s Existing Contributors?
Existing contributors face a clear adjustment: every pull request must now be human-authored and human-understood. According to PC Gamer’s coverage, the Godot Foundation stated that maintainers “can’t trust heavy users of AI to understand their code enough to fix it” (PC Gamer, 2026). This directly impacts developers who previously relied on Copilot or similar assistants for routine patches.
The policy shift means contributors must verify they fully comprehend each line they submit. Maintainers have reported increasing difficulty reviewing submissions where the author cannot explain their own logic. This creates review bottlenecks. The foundation wants to eliminate that friction entirely.
For long-time contributors who write code manually, the change is effectively transparent. The ban targets AI-authored submissions, not developers who happen to use AI for unrelated tasks. However, anyone who integrated AI into their Godot workflow needs to adapt. The expectation is personal accountability for every contribution.
Are Documentation and Comments Also Covered by the Ban?
Yes, the ban extends beyond code. According to DualShockers, the policy covers “code, agents, and AI-generated text in human comms” (DualShockers, 2026). This means documentation, commit messages, pull request descriptions, and code comments all fall under the restriction.
The reasoning is practical. Maintainers review documentation and communication to understand a contributor’s intent and technical grasp. When AI generates these materials, the reviewer loses signal about the contributor’s actual understanding. AI-written PR descriptions often sound polished but obscure whether the author truly grasps the problem. The foundation treats this as a review integrity issue.
Contributors should write their own commit messages and documentation. The policy does not prohibit spell-checkers or grammar tools. It targets substantive AI generation. If a contributor cannot describe their own change in plain language, that signals a problem the foundation wants addressed before review begins.
What Tools and Agents Are Explicitly Named in the Policy?
The policy does not single out specific products by name. Instead, it categorically covers AI code agents and AI-authored text generation tools. DualShockers reports the ban explicitly includes “code, agents, and AI-generated text in human comms” (DualShockers, 2026). This sweeping language covers the full spectrum of current tools.
That means GitHub Copilot, Claude, ChatGPT, Gemini, Cursor, and autonomous coding agents all fall within scope. The foundation chose broad language deliberately. Naming individual tools would require constant updates as new products launch. A categorical ban avoids that maintenance burden.
The policy also covers emerging autonomous agents that write and submit code independently. These tools have become more prevalent in open-source ecosystems. By banning agents specifically, Godot closes the door on automated PR generation entirely. No AI tool gets a pass.
How Does Godot’s Stance Compare to Other Open-Source Projects?
Godot joins a growing list of open-source projects pushing back against AI contributions. The policy is notably stricter than many projects that allow AI-assisted code with disclosure. According to GamingOnLinux, Godot is “adjusting their policies to get stricter on AI code contributions” (GamingOnLinux, 2026). The total ban sets Godot apart from projects with more permissive stances.
Some open-source communities require contributors to label AI-assisted pull requests. Others prohibit AI for security-sensitive modules but allow it elsewhere. Godot’s blanket prohibition removes ambiguity. There is no disclosure workaround.
The divergence reflects differing community needs. Projects with large review teams sometimes accommodate AI-assisted submissions. Smaller maintainer teams like Godot’s face review capacity constraints. AI slop compounds that problem. The total ban is a direct response to operational pressure.
What Should Developers Do Before Submitting a Godot PR?
Developers should write every line of code, comments, and documentation manually before submitting. They must be prepared to explain their logic, answer technical questions, and fix issues without AI assistance. PC Gamer reports the foundation’s position: contributors must understand their code well enough to fix it themselves (PC Gamer, 2026). This is the core expectation.
Before opening a pull request, developers should review their diff carefully. Remove any AI-generated boilerplate. Rewrite commit messages in their own words. Ensure the PR description accurately reflects their personal understanding of the change.
If a contributor used AI to explore approaches during development, that exploration phase is not banned. The policy targets what enters the repository. The final submission must be human-authored and human-understood. When in doubt, contributors should ask maintainers on GitHub or the Godot contributor chat before proceeding.
Frequently Asked Questions
Can I still use AI tools like Copilot for personal Godot projects?
Yes. The ban applies exclusively to code submitted to the official Godot Engine repository. Personal projects built with Godot are unaffected. According to PC Gamer, the policy governs contributions to the engine itself, not games or applications created with it (PC Gamer, 2026). Developers retain full freedom to use AI tools in their own Godot-based projects.
Will the Godot project accept AI-generated bug reports?
No. The ban covers AI-generated text in human communications, which includes bug reports and issue comments. DualShockers confirms the policy covers “AI-generated text in human comms” broadly (DualShockers, 2026). Bug reports must be written by humans who can reproduce and explain the issue directly.
Does the new policy apply to Godot Engine asset library submissions?
The Godot Foundation’s policy targets the core engine repository and related communication channels. Asset library submissions were not specifically carved out in the announcement. GamingOnLinux reports the policy addresses AI code contributions to the engine project (GamingOnLinux, 2026). Contributors should check with asset library maintainers for project-specific guidelines.
What happens if a contributor unknowingly submits AI-generated code?
The foundation expects contributors to verify their submissions are human-authored before opening a PR. PC Gamer’s coverage emphasizes that maintainers “can’t trust heavy users of AI to understand their code enough to fix it” (PC Gamer, 2026). Unintentional AI-generated submissions risk rejection. Maintainers will evaluate each case individually.
Summary
Godot’s ban on AI-authored contributions reflects three key realities for the open-source ecosystem:
- Accountability is non-negotiable. Contributors must understand and fix their own code. AI tools cannot take responsibility for bugs, and contributors using them heavily often cannot either.
- The ban is comprehensive. Code, documentation, commit messages, PR descriptions, and bug reports all fall under the policy. No AI tool or agent is exempt.
- Review capacity drives the decision. Maintainer teams lack the bandwidth to evaluate submissions where the author cannot explain their own logic. The ban is a practical response to operational constraints.
- Personal projects remain unaffected. Developers can use AI freely in their own Godot-based games and applications. The policy governs only contributions to the engine itself.
If you contribute to Godot or plan to, read the full policy on the project’s GitHub repository. Write your own code, write your own communications, and be ready to defend every line you submit. The era of AI slop PRs in Godot is over.