This covenant sets the norms for responsible AI use in the LOKF project. LOKF is developed openly and with substantial AI assistance; this document makes the resulting expectations explicit: AI is a tool that accelerates the work, but every contribution is owned, understood, and defended by a human.
It applies to this repository:
It is adapted from the LinkML AI Covenant.
Everything you contribute is yours — regardless of what tools helped create it.
When you submit code, schema changes, documentation, issues, or comments with AI assistance, you are the author. You are responsible for:
- Understanding what you are submitting
- Verifying its correctness and appropriateness
- Defending and explaining your choices during review
- Ensuring it meets the project's standards
Do not submit anything you cannot fully stand behind.
AI tools are not credited as authors or co-authors. Commit messages and pull
requests MUST NOT add Co-Authored-By trailers (or equivalent) for an AI
assistant. The human who runs the tool is the sole author and takes full
responsibility for the result.
LOKF is defined once in lokf.yaml; the JSON-LD context, JSON Schema, SHACL
shapes, and OWL ontology are generated from it. AI assistance does not change
that contract: propose changes in lokf.yaml (or the toolkit, docs, or tests),
regenerate with just build, and commit the regenerated artifacts — never
hand-edit a generated file, whether by hand or with an AI.
AI review tools (Claude, Copilot, CodeRabbit, etc.) provide automated quality checks, not human reviews.
- AI comments are suggestions, not requirements.
- PR owners may resolve AI comments without responding.
- Human reviewers may use AI feedback to inform their own review.
- A PR still requires human approval regardless of AI feedback.
AI tools can be helpful thinking aids when preparing to participate in issues and discussions. They may be used to clarify your own thinking, explore alternative framings, or help draft your contribution for clarity and structure.
However, discussions exist to surface, negotiate, and consolidate human judgement. AI systems MUST NOT be used to autonomously post comments, replies, or messages in GitHub issues or discussions. Every contribution must reflect a human position the author is prepared to explain, revise, and defend. Posting AI-generated commentary as an independent "voice" undermines trust, accountability, and the purpose of deliberation.
In short: AI may support participation; humans must own it.
Required:
- When proposing a fix or change to code or schema you don't fully understand, attribute the idea to AI so reviewers can assess it appropriately.
Appreciated:
- When brainstorming, distinguish "AI suggests X" from "I recommend X based on my own judgement." This helps prioritize ideas.
Not required:
- Routine use of AI to write code, schema, docs, issues, or PR descriptions.
- (AI co-authorship in commit messages is not merely "not required" — it is disallowed; see No AI co-authorship.)
| Situation | Guidance |
|---|---|
| Writing code/schema/docs with Claude/Copilot | No disclosure needed; you own the result |
| Submitting an AI-suggested change you fully understand | No disclosure needed |
| Submitting an AI-suggested change in unfamiliar territory | Disclose the AI origin for reviewer context |
| Drafting an issue or PR description with AI | No disclosure needed; ensure it's accurate |
| Brainstorming in discussions | Be clear about AI-generated vs. your own ideas |
| Receiving AI review comments | Address or resolve at your discretion |
| Crediting an AI as a commit author or co-author | Not allowed |
This covenant is built on trust. By contributing to LOKF, you agree that:
- You will not submit AI-generated content without reviewing it.
- You take responsibility for any issues arising from your contributions.
- You will be honest about the origin of ideas when it matters for review quality.
This covenant may evolve as AI tools and community needs change. It is adapted from the LinkML AI Covenant. Feedback and suggestions are welcome.