Add an explicit AI policy to our documentation

[pfmoore]

As noted in #13825, we're starting to get AI-generated contributions. In order to ensure our limited maintainer bandwidth isn't overwhelmed, and to protect us from any ownership or licensing concerns over submitted code, I suggest that we explicitly adopt and document an AI policy. My strong preference is for something based on the attrs policy - https://github.com/python-attrs/attrs/blob/main/.github/AI_POLICY.md. In my view, this takes the correct stance, focusing on ensuring that every PR has a human contributor associated, and that contributor takes full resaponsibility for the technical and legal/ownership aspects of their contribution.

There are always going to be judgement calls involved - after all, LLMs by design produce plausible looking content - so the policy will of necessity be subject to interpretation. By choosing a policy that assumes good faith on the part of submitters, while still being clear that we will impose penalties like rejection or banning when necessary, I think we strike a reasonable balance.

I have created #13826 as a starting point for discussion.

[notatallshaw]

+1 to a policy like this, it makes sense to outline our policy publicly, I'll try and take some time to review the wording soon.

I would love to see it be more concise and make clear that directly replying with AI/LLM responses anywhere, not just in PRs, is not appreciated and may lead to be marked as spam or banned. I am also concerned about some of the more definitive statements about copyright it is making but that's because I'm not an expert on the legal side, and so would appreciate other maintainers reviewing that.

I still think from a practical point of view it makes sense to add an entry point for agents as I outline in #13825, but I think one of the big benefits it could have is instructing a bot that the user confirm they have read the policy before contributing, so let's work on this first.