Skip to content

AI_POLICY.md

Latest commit

 

History

History
69 lines (45 loc) · 7.39 KB

AI_POLICY.md

File metadata and controls

69 lines (45 loc) · 7.39 KB

AI Contribution Policy for the Strimzi Project

Overview

Contributors may use AI tools, such as Large Language Models (LLMs) and code assistants, when contributing to Strimzi. As with any development tool, the contributor is responsible for the quality of the result and for understanding what they submit.

You are the contributor. When you submit a Pull Request (PR) to Strimzi, you certify the contribution as your own work. AI-generated or AI-assisted content does not change this responsibility.

Requirements

Understanding and Responsibility

  • You must understand your contribution. Do not submit code, documentation, or other content that you do not fully understand. You should be able to explain to reviewers what your contribution does, how it works, and how you verified it. You should be able to answer reviewers' questions yourself, without recourse to AI. In particular, do not waste the time of other contributors by being a proxy between reviewers and an AI.
  • Review all AI-generated content. Before submitting, verify that the contribution is correct, tested, and follows the project's contributing guidelines.
  • You own what you submit. As the PR author, you are responsible for all changes, regardless of the tools used to create them, and must certify that you wrote or otherwise have the right to submit the code you are contributing by signing your commits.

Disclosure

  • Disclose AI usage in your PR description. If you used AI tools in preparing your PR, you must disclose this in the description of your PR. Listing AI tooling as a co-author, co-signing commits using an AI tool, or using the Assisted-by, Co-authored-by, Co-developed-by, or similar commit trailer is not allowed.

What NOT to Do

  • Do not use AI tools in commit trailers. AI tools MUST NOT be added to Signed-off-by, Co-authored-by, Assisted-by, Co-developed-by, or similar tags in commit messages. Only humans can legally certify the Developer Certificate of Origin (DCO). The Signed-off-by tag certifies that you wrote the code or otherwise have the right to submit it under the project's license. An AI tool is not a legal entity and cannot make this certification.
  • Do not instruct AI tools to directly respond to comments. When maintainers, component owners, or other contributors provide feedback on your PR, you may use AI tools to help craft responses, but you must fully understand and take ownership of what you write. Reviewers want to engage with you, not receive generic AI-generated responses. If you use AI assistance to formulate responses, ensure they genuinely reflect your understanding and that you can defend your reasoning.
  • Do not submit large AI-generated PRs. AI tools can generate content faster than reviewers can read it. Keep PRs focused and appropriately sized. Strimzi maintainers MAY close pull requests that are too large to review effectively or where the contributor does not appear to understand the changes.
  • Do not open many AI-generated PRs in rapid succession. Submitting many PRs at once or within a few minutes overwhelms reviewers and will result in PRs being closed. Instead, focus on quality over quantity.
  • Avoid verbose AI-generated content. Trim down unnecessary verbosity in PR descriptions, commit messages, and code comments. Be clear, concise, and specific. Respect the time of maintainers and reviewers.
  • Do not use AI for good-start issues. Issues labeled as good-start are intended for new contributors to learn the codebase and gain experience. Using AI tools to complete these issues is against their purpose and takes learning opportunities away from newcomers.

Quality Standards

  • Meet the same standards. AI-assisted contributions are reviewed to the same standard as any other contribution. Code must be correct, maintainable, well-tested, and consistent with project conventions.
  • Ensure licensing compliance. AI-generated content must not introduce material under licenses incompatible with the project's Apache 2.0 License.
  • Test your changes. All code contributions must be tested according to the project's standards, whether AI-assisted or not.

Consequences of Non-Compliance

Pull Requests that violate this policy may be closed. Specific cases include:

  • PRs where the contributor cannot explain or defend the changes during review.
  • PRs with generic, inauthentic responses to reviewer feedback that demonstrate a lack of understanding.
  • Large, unfocused PRs that appear to be bulk AI-generated content.

AI-Assisted Pull Request Reviews

Automated or AI-assisted reviews can supplement the review process but do not replace review by maintainers or component owners.

  • AI reviews don't replace human review. Automated checks and AI-generated review comments can provide useful signals (e.g., identifying security vulnerabilities, style issues, or potential bugs), but they do not satisfy the project's review requirements.
  • Humans make the decisions. Maintainers and component owners make merge decisions following the project's governance framework. PR approval requires votes from maintainers, component owners, or both as defined in the governance policy.
  • Contributors should engage with human reviewers. While you may find AI-assisted review tools helpful during development, your PR must be reviewed and approved by project maintainers or component owners before merging.
  • Contributors should not request AI-assisted reviews on PRs. While you may use AI tools to review your code privately during development, do not invoke AI bots or tools to post review comments on project PRs. If AI-assisted review is desired, maintainers will request it.

AI-Assisted Strimzi Proposals

Strimzi proposals are strategic documents that significantly impact users or project direction. Due to their importance, special considerations apply:

  • AI-assisted proposals require deep understanding. If you use AI tools to help draft a proposal, you must have a thorough understanding of the problem space, proposed solution, alternatives considered, and implications for users and the project.
  • The rationale matters most. Proposals are not just about "what" to do, but "why" it should be done, what alternatives exist, and what trade-offs are being made. AI tools may help structure these ideas, but the reasoning must come from genuine understanding of the project and its users.
  • Disclose AI usage. If AI tools play a significant role in drafting a proposal, note this in the PR description.
  • Engage authentically in discussions. Proposal discussions involve community consensus-building and voting. You must engage directly with the community, not through AI-generated responses. Maintainers and community members need to understand your reasoning and evaluate the proposal based on your understanding, not an AI-generated output.
  • AI-assisted reviews of proposals. You may use AI tools to help you understand and evaluate proposals, but your vote or feedback must reflect your own judgment and reasoning.

Proposals that appear to be largely AI-generated without genuine understanding of the problem space will be closed by maintainers.

Questions

If you have questions about this policy or are unsure whether your use of AI tools requires disclosure, please contact the maintainers via Slack or the mailing list.