Update code of conduct to include guidelines around LLM usage

[nzakas]

On ESLint, we've recently started seeing an uptick of people submitting issues and PRs that are entirely AI-generated. In many cases, people have clearly not even reviewed the content that was generated by the LLM before submitting and in one case the user kept feeding our replies back into the LLM to generate comments. I fear that this trend is going to get worse going forward.

As such, I'd like to request an addition be made to the Open JS Code of Conduct to reflect AI usage. I'd like to see something along the lines of (very, very rough):

• We allow issues and PRs created by LLMs
• You must transparently indicate when issues and PRs are created by LLMs
• We expect that you have actually reviewed what the LLMs created before submitting them
• When maintainers interact with you, your responses must not be created by LLMs

FWIW, Preact recently updated their CoC for this very purpose.
preactjs/preact#4961

[ljharb]

TC39 has been having discussions about this very topic, with most opinions being that the CoC isn't the appropriate place for this - contributor guidelines might be a better fit?

[tobie]

See #1509.

[nzakas]

I think the CoC is an appropriate place to mention this as it relates to contributor behavior and could potentially lead to the banning of someone from a repo.

[tobie]

imho, more than a CoC issue, this is a moderation issue. AI slop is essentially spam. Projects are already empowered to moderate their repositories as they see fit.

[ljharb]

CoCs aren't about all behavior, though - and bans haven't ever solely been due to CoC violations, they can be for any reason at any time.

[nzakas]

Regardless of what projects are allowed to do now, I still think having some coherent policy about AI-generated contributions would be helpful rather than asking each project to come up with their own approach.

[ljharb]

I completely agree with that! The AI Collab Space is probably the right place to come up with something.