add ai policy to contributing.md - #4835
Conversation
YoshiRulz
added
the
Meta
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
This comment was marked as outdated.
This comment was marked as outdated.
Empty request for changes is disruptive.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
|
There. |
|
Added checkboxes to issues and PRs for good measure. Requiring disclosure is an iffy subject because it may make it harder to distinguish blatant violations from people who properly used AI. If they haven't used it at all, they will check the box. If they used it in some way, only having the blanket checkbox would attempt to force them to describe in human text how it was involved. And if it's silently ignored or removed, we'll know it's suspicious and ask direct questions. |
|
|
||
| Submoduling third-party code that involves output of the generative AI is allowed if there is no viable alternative free from generative AI. | ||
|
|
||
| Using generative AI as a tool to aid with your research prior to submitting content to this repository is only allowed if that content does not include any AI hallucinations: bug reports and proposed fixes must be valid and sensible. |
There was a problem hiding this comment.
Do we want people to state in issues or pull requests when they use generative AI in this way?
There was a problem hiding this comment.
I think the wording of the checbox solves it, as explained in my post above.
There was a problem hiding this comment.
I disagree. Intentionally leaving out information about what you want from other people should not be expected to result in other people doing what you want, nor should it be expected to encourage good communication.
There was a problem hiding this comment.
What should be expected then?
There was a problem hiding this comment.
Disappointment, frustration, and/or disengagement.
There was a problem hiding this comment.
Not seeing it, please make an example scenario.
|
|
||
| ## Generative AI policy | ||
|
|
||
| Content created by generative AI, including large language models (LLMs), should not be submitted to this repository. That involves text and media files, issues and pull requests. |
There was a problem hiding this comment.
Is AI translation permitted?
There was a problem hiding this comment.
Translation is technically not generation of thoughts that are actually being conveyed. Unlike tasvideos the site, we don't have to go out of our way to support non-native speakers. But if it's used, we could tolerate that. I don't feel it needs a mention in the policy, let it remain vague, to encourage clarification.
There was a problem hiding this comment.
I don't like vague policies.
+If you are not fluent in English, you may use machine translation when submitting an issue or PR, so long as your original text is included.
There was a problem hiding this comment.
When adding a line such as this, it often also helps to mention something along the lines of
+Your English does not need to be perfect. We prefer slightly broken English over machine translations. However, if that doesn't suffice, you may use machine translation when submitting an issue or PR, so long as your original text is included.Either that, or don't mention it at all, as to not entice people to use machine translations when not needed.
Co-authored-by: YoshiRulz <OSSYoshiRulz+GitHub@gmail.com>
5 participants
Text very loosely based on https://github.com/libsdl-org/SDL/blob/main/AGENTS.md