Skip to content

AI_POLICY.md

Latest commit

 

History

History
81 lines (59 loc) · 4.48 KB

AI_POLICY.md

File metadata and controls

81 lines (59 loc) · 4.48 KB

AI Policy

Summary

Contributions generated by AI or with AI assistance are welcome, but strict rules apply.

These policies are zero tolerance policies, meaning we reserve the right to ban or otherwise remove from participation anyone caught breaking them.

  • It is never acceptable to directly copy AI responses as if they were your own words! This differs from our policy on code contributions, which you accept full responsibility for.
  • All work is the responsibility of the human who makes the PR. We do not accept AI as co-contributors on commits.
  • All AI contributions must be fully vetted by the submitter.
  • AI can produce a lot of text. Just because you can write a 500 line github issue does not mean our team has time to read it. Be prudent.
  • We offer guidelines and prompting advice on how to most effectively use AI, but do not encourage or require its use. The tools you use are your own business.
  • If the project you are contributing to has an AGENTS.md file, it is expected that you use it in your agent context. The presence of an AGENTS.md file does not explicitly condone or request AI usage.
  • Security and vulnerability reports are special cases whereby an LLM must not be involved in essentially any way.

Your Own Voice

Of all of the rules, this is the one that is most likely to get you banned for violating. One major exception to this rule is that using LLMs to translate is perfectly acceptable and welcome. We are an international community and welcome all.

It is never acceptable to copy the response of an LLM or AI tool and use it as if it were your own words. It is acceptable to link to an agent conversation, or to clearly annotate some piece of information as generated by AI, but it is generally discouraged. If we want to talk to an AI, we will ask them ourselves. We expect all interactions to be human interactions.

This applies to:

  • Comments
  • Pull Requests, excluding parts explicitly stated as LLM generated
  • Issues, excluding parts explicitly stated as LLM generated

This does not apply to:

  • Code
  • Documentation
  • Commit Messages

Personal Responsibility

Any contributions made are yours and yours alone. You are accepting responsibility for them by making a pull request or opening an issue. As such, we do not accept AI as co-contributors on commits. In addition, it is unacceptable to submit un-reviewed AI code. All code, human-generated or otherwise must be fully vetted and vouched for by you. If we detect that this is not the case, you may be banned. This means there is an inherent risk to making AI generated PRs, as they may even be held to a higher standard than normal. Please keep this in mind.

Spam

While spam in any form is unacceptable, it can be very easy to accidentally spam with AI. AI tools can generate massive amounts of text and code. Remember that humans have to read what you are generating. Less is more.

When opening Github issues, for example, AI often generates overly long-winded and assumption-filled issues and proposals. Stick to the essential and verifiable information only. See the guidelines below for advice on prompting your agent if you must use one to generate an issue or proposal description.

Security & Vulnerability Reports

Security & vulnerability reports are a special case to this policy because of their increased demand for urgent attention from our team. These issues/reports must be reproduced, written, verified, and analyzed by a human being. We simply cannot allow there to be a machine-to-team pipeline for high-priority security-related issues.

Guidelines

  • If you are making pull requests, make the smallest reasonable pull request possible. This is true regardless of AI usage, but especially necessary to keep in mind in the context of AI, otherwise you risk overloading our team.
  • When opening an issue, provide the appropriate issue template to your agent.
  • When opening a pull request, provide the pull request template to your agent.
  • When in doubt, copy this entire document into your prompt and ask the agent what you should do 🙂
  • If the project you are contributing to has an AGENTS.md file, it is expected that you use it in your agent context.
  • Encourage agents to be as concise as possible and not to make any assumptions.

Derivative Works

Anyone is free to copy and/or adapt this in any way.
⚡