DOC: Refine AI policy

- Added "human" in
  > the resulting contribution is the responsibility of the human contributor
  to clarify the human oversight. I recently had the feeling that AI agents would think they are compliant without a human because they understand themselves as contributor.
- Explicitly listed "Simultaneously contributing to many projects." as not acceptable. This "shotgun contribution" typically lacks human oversight, and even if somebody manages to do this with high quality, the FOSS community as a whole would be better of if they instead contributed more in-depth to one or two projects.

Author: timhoffm

doc/devel/contribute.rst

@@ -189,7 +189,7 @@ Use of Generative AI
 ====================
 
 Generative AI tools are evolving rapidly and can be helpful. As with any tool,
-the resulting contribution is the responsibility of the contributor. We
+the resulting contribution is the responsibility of the human contributor. We
 expect dedicated and authentic engagement in our community. In particular when
 using AI, carefully consider what and how to communicate, question results,
 think things through thoroughly and make well-informed decisions.
@@ -216,6 +216,9 @@ Some examples of acceptable and unacceptable AI uses are:
     - Solving topics that you wouldn't be able to solve yourself without AI
     - Using AI output without ensuring that you fully understand the output or
       without verifying that it is the correct approach.
+    - Increase breadth of contributions, i.e. imultaneously contributing to several
+      projects. Instead of spreading your resources, you can provide greater value
+      by engaging more deeply with one or two projects.
 
 To ensure project health and preserve limited core developer capacity, we will flag
 and reject low-value contributions that we believe are AI generated. We may ban