Define guidelines & processes for LLM/AI-based contributions

[david-allison]

I'm opening this issue for discussion. There have been a few comments on Discord (and the GSoC Mentor Summit/Mailing List). The burden on reviewers is steadily increasing, but the meaningful benefit from LLM-generated Pull Requests doesn't justify this additional burden.

The aim is to:

• Define our current problems
• Research what other orgs are doing for these problems
• Brainstorm reasonable solutions

Outcomes

I don't want to be heavy-handed here, a few incremental changes will likely suffice. I suspect that this will result in some of:

• An update to our PR Template
• An 'acceptable use' policy for LLM-based contributions
• Potential configuration changes to GitHub Copilot
• An update to CONTRIBUTING.md
  • on top of https://github.com/ankidroid/Anki-Android/pull/19732
• CI: Validate the PR Template is filled in

Past AnkiDroid Discussions (Discord)

I will summarize these later:

• https://discord.com/channels/368267295601983490/701922522836369498/1436535978037608518
• https://discord.com/channels/368267295601983490/701922522836369498/1447289676598611978
• https://discord.com/channels/368267295601983490/701922522836369498/1442784489909649479