Repository navigation
Conversation
|
@github-actions crossbow submit preview-docs |
|
|
This comment was marked as outdated.
This comment was marked as outdated.
|
@github-actions crossbow submit preview-docs |
This comment was marked as outdated.
This comment was marked as outdated.
There was a problem hiding this comment.
🟢 Approval recommended
The documentation changes address the requested reviewer guidance with no unresolved blocking issues.
No open issues 🎉
What changed in this PR
This documentation-only PR extends Arrow’s AI contribution guidelines with expectations for human participation during review, addressing #55259.
Changes:
- Clarifies that contributors should engage directly with reviewers.
- Documents possible closure or labeling of PRs showing excessive reliance on AI.
- Adds an example reviewer message linking to the AI-generated code policy.
| File | Description |
|---|---|
docs/source/developers/overview.rst |
Adds human-engagement expectations and guidance for responding to excessive AI usage. |
Review effort: Balanced
AlenkaF
left a comment
AlenkaF
left a comment
There was a problem hiding this comment.
Thanks! Just one nit connected to the dropdown box.
Co-authored-by: Alenka Frim <AlenkaF@users.noreply.github.com>
|
@github-actions crossbow submit preview-docs |
|
Revision: 30347ce Submitted crossbow builds: ursacomputing/crossbow @ actions-be29f0bb4a
|
|
I'm not sure the added paragraph adds a lot of value. All that is being said there is already said above in one form or the other. For example:
was already expressed in:
|
|
IOW, you might want to reword the existing test instead of adding more. Adding more text is not useful in itself. |
| and reviewers are not expected to be talking to contributors' agents but to | ||
| contributors themselves. If a reviewer finds that a contributor seems to rely | ||
| excessively on AI tools instead of engaging with the review, the PR might be | ||
| closed and/or tagged with a label indicating excessive AI usage. |
There was a problem hiding this comment.
I appreciate we have a need for filtering out low-effort PRs, but I worry such a tag would create unnecessary drama, and think we might have an easier life by creating a positive tag and recommending reviewers filter on that if they want a low chance of AI PRs
For what could qualify someone for a positive tag, I quite like the prioritisation order listed at the bottom of this blog post by Andrew Lamb: https://andrew.nerdnetworks.org/blog/2026/10/19/my-plan-for-reviewing-prs-in-open-source-projects-in-the-age-of-ai/
There was a problem hiding this comment.
Actually arrow-rs auto labels "Prolific Author" - https://github.com/apache/arrow-rs/pulls?q=is%3Apr+label%3A%22Prolific+Author%22
There was a problem hiding this comment.
Sure, but that's something else entirely, just by the same person.
There was a problem hiding this comment.
FWIW I'm opposed to that as well as it's labelling the individual not the action. It's going to feel very personal to anyone who gets a false positive.
We could achieve the same thing by just closing those PRs with an explanatory message, if we wanted to limit PRs from people who have made lots of PRs to lots of repos.
There was a problem hiding this comment.
(for clarity my original comment was referring only to "tagged with a label indicating excessive AI usage" and not the rest of the change)
github-actions
Bot
added
awaiting changes
| .. dropdown:: Example of message for a PR with excessive AI usage | ||
| :animate: fade-in-slide-down | ||
| :class-container: sd-shadow-none | ||
|
|
||
| This PR doesn't appear to comply with our policy on tool-generated content. | ||
| Please see our AI generated code policy on AI-generated code: | ||
| https://arrow.apache.org/docs/developers/overview.html#ai-generated-code | ||
|
|
There was a problem hiding this comment.
I like that we are providing an example for folks to use. I wonder if we want a reviewer's guide to put this into perhaps?
Rationale for this change
Add documentation of our practices for AI generated code/review comments.
What changes are included in this PR?
Some documentation with a proposed message and indications about possibility of closing PRs or adding labels indicating this might be AI-slop or requires a human in the loop.
Are these changes tested?
No only docs.
Are there any user-facing changes?
No
Was AI used for this PR?
In accordance to the AI generation guidelines, please disclose below whether and how AI was used in this PR.
PR code and description written by:
Reviewed before submission by: