Add AI contribution policy section to the README.md

This is a verbatim copy of the one from GTK at
https://gitlab.gnome.org/GNOME/gtk/-/blob/main/CONTRIBUTING.md#ai-contribution-policy

This might be fine-tuned further in the future.

Author: sdroege

README.md

@@ -34,6 +34,55 @@ When using crates that are not part of the `gtk-rs` repository, you will
 need to be careful and ensure that they do not pull in incompatible versions of core
 crates like `glib-rs`.
 
+## AI Contribution Policy
+
+gtk-rs is a project by humans for humans. We prefer contributions that
+are produced by human creativity, we expect a human to take full
+responsibility for each contribution, and we will take more joy in
+reviewing contributions when there's people at the other end of the
+line to stand by their changes.
+
+If you use LLM/GenAI tools for your contributions, here are the rules
+you must follow:
+
+### Requirements
+
+1. Use AI as a tool. Verify behavior, correctness, and compatibility
+   yourself prior to submitting your contribution. Do not ask the
+   maintainers to do this for you.
+1. Keep changes narrow and limited. Do **NOT** use LLM/GenAI tools to
+   generate broad rewrites, large refactorings, or style changes.
+1. Do **NOT** submit generated code, documentation, or tests that you
+   don’t understand.
+1. Do **NOT** fabricate benchmarks, bug reports, test results, code
+   samples, or reproducers.
+1. Do **NOT** include private code, credentials, tokens, or any other
+   confidential material.
+1. Respect the licensing and attribution requirements.
+
+### Disclosure
+
+Always disclose the use of LLM/GenAI tools when creating an issue or
+a merge request. Do not include trailers like “Co-authored-by:” or
+“Assisted-by:” in commit messages, since they serve as free advertising
+for AI companies.
+
+### Reviews
+
+1. Describe your changes, and the verification steps.
+1. Be prepared to explain all the changes yourself.
+1. Do **NOT** feed the review feedback to an LLM/GenAI tool.
+
+### Maintainers expectations
+
+1. Review LLM/GenAI-assisted contributions more strictly than any other contribution.
+1. Require reproducibility in fixes and tests.
+1. Reject changes that appear to be unverified LLM/GenAI output.
+1. Reject comments and feedback that appear to be LLM/GenAI output.
+
+> A COMPUTER CAN NEVER BE HELD ACCOUNTABLE.
+> THEREFORE A COMPUTER MUST NEVER MAKE A MAINTENANCE DECISION.
+
 ## Regenerating
 
 To regenerate crates using [gir], please use the `generator.py` file as follows: