NNCF welcomes responsible use of AI tools in open-source collaboration. This policy exists to protect maintainer time, keep review quality high, and ensure long-term maintainability of the project.
The key principle is simple: contributions are evaluated by quality, accountability, and maintainability.
- Pull requests and code changes
- Issues and discussions
- Review communication on GitHub
- Read and review all generated code or text before asking other project members to review it.
- Ensure that all contributions comply with the contribution guide.
- Understand your submission end-to-end and be ready to explain design and implementation decisions.
- Verify correctness yourself (build, tests, behavior, and edge cases).
- Take full responsibility for every line submitted, regardless of how it was drafted.
The following are not acceptable and may lead to immediate closure of the contribution:
- Submitting code you cannot explain or maintain
- Large, low-context, or low-quality AI-generated changes without thorough human validation
- Using AI-generated responses in place of direct, human-to-human communication during review
- Auto-generated issues/discussions that do not describe reproducible, first-hand observations
- Fabricated citations, benchmarks, bug reports, or security claims