AI / LLM usage for code and communication is not completely prohibited, but please don't. There are 2 main reasons:
- Time: Generating large volumes of content with LLMs is simple, but reading and trying to understand it takes time. Generated text usually is needlessly verbose, and no one can tell which parts you actually care about and what was generated in addition.
- Respect: When someone volunteers to work on your request, they want to deal with an actual human who reacts to feedback.
Effectively LLMs increase the maintainer workload when there already is too much to do, and it's no fun at all to deal with. No fun may sound irrelevant, but in fact can be very destructive for a hobby project such as HeliBoard.
- Communication must happen between humans. You may use AI for translations or similar, but not for generating comments.
- Contributions must be made with human interaction, i.e. AI generated PRs and issues are prohibited.
- Understand your contributions. AI assistance is acceptable and must be disclosed, but you still should be able to understand and explain the code yourself.
- Your contributions need to be compatible with the project's license. This effectively means an LLM can only be trained on code compatible with GPL3.