Update AI_POLICY.md

[orbeckst]

Changes made in this Pull Request:

• make this policy the default in the MDAnalysis org
• allow separate policies on other repos/projects

Once this PR is merged, we can link to this policy from the org README.

LLM / AI generated code disclosure

LLMs or other AI-powered tools (beyond simple IDE use cases) were used in this contribution: no

PR Checklist

• Issue raised/referenced?
• Tests updated/added?
• Documentation updated/added?
• package/CHANGELOG file updated?
• Is your name in package/AUTHORS? (If it is not, add it!)
• LLM/AI disclosure was updated.

Developers Certificate of Origin

I certify that I can submit this code contribution as described in the Developer Certificate of Origin, under the MDAnalysis LICENSE.

---

📚 Documentation preview 📚: https://mdanalysis--5391.org.readthedocs.build/en/5391/

[orbeckst]

This PR follows from a conversation in #develop https://discord.com/channels/807348386012987462/808088023957897258/1509685572288118784

@orbeckst : @MDAnalysis/coredevs should we include a link to our org-wide AI Policy in the org-level README at https://github.com/MDAnalysis ?

@IAlibay : Sounds reasonable, we might want to update the policy to say something like "individual sub-projects may define their own policies". I'm thinking there's a few projects the coredevs don't really speak for that are in the org.

[orbeckst]

Note that at the moment the MDAnalysis/mdanalysis AI Policy effectively applies to any project/repo in our org. The proposed changes allow the people in charge of each repo to adopt their own AI policy. The proposed change to our AI Policy contains no restrictions on what these individual AI Policies may contain. If we as CoreDevs want to provide any guidance then we have to add it explicitly. Of course we always have the nuclear option of moving a project out of the org if we find that it does no longer aligns with our mission and values.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.85%. Comparing base (523c853) to head (8aae559).

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #5391   +/-   ##
========================================
  Coverage    93.85%   93.85%           
========================================
  Files          182      182           
  Lines        22509    22509           
  Branches      3202     3202           
========================================
  Hits         21125    21125           
  Misses         922      922           
  Partials       462      462           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[orbeckst]

My own view is to rather err on the side of giving more freedom and trusting that people make sensible choices than to be overly prescriptive.

The risks that I see are

1. The license of a code base is being polluted and invalidated/changed and that may be relatively difficult to fix.
2. Contributors may get confused when contributing to repos with starkly differing policies.
3. Our position as an organization on AI questions may become less credible if we are seen to support very different approaches to AI in open source.

[IAlibay]

Our position as an organization on AI questions may become less credible if we are seen to support very different approaches to AI in open source.

I don't think this is a problem. I suspect the AI policy may change in downstream tools more than upstream (because upstream tools are handled by coredevs and we're always careful not to pollute the license). So at worst the argument is going to be that libraries like MDAnalysis require stricter AI policies than applications / MDAKits.