As the title notes add an option when flagging to note that the post is/appears AI generated.
The flag should note if the site allows AI generated content or not so the creator of the post knows. If AI content is allowed the flag should note what is needed, e.g. the model and version of the model used, and possibly the prompt used to create the generated content.
On the OpenAI site there was a DevDay a few weeks ago and the OpenAI Discourse forum grew in number of post, then the CEO being fired and a change of the board of directors brought in more new users. With that we are now seeing many first post in a topic that are just AI generated. While the regular users are still trying to figure out if these should be allowed, having the option to at least flag these so that other users know the post is AI generated content would be of value.
Yes I know there is the something else option when flagging. However many users do not use it and often if a user uses it one has to be a mind reader to understand why the user really is flagging. The description for something else are often so cryptic that a moderator has to DM the user to get an understanding of the flag. A flag option that is in the list is often quickly picked and used.
I know some may consider this a controversial feature request so expect either a lot of discussion or next to no discussion.
If you do want to discuss this please be nice, stay on topic, no attacks against an individual and as always abide by the site guidelines.
This seems like a good candidate for a site-defined custom flag. I know there have been a few requests for that, though I’m not sure it’s on the immediate roadmap.
Excerpt: “This seems like a good candidate for a site-defined custom flag. I know there have been a few requests for that, though I’m not sure it’s on the immediate roadmap.”
I was looking for a good feature topic to link to, but I think a large chunk of recent discussion about it has been internal. Still, it has gained some traction and I think this is another good use case for it.
I’ve removed the ai plugin tag as it’s more to do with moderating ai content rather than a feature of the plugin. Though I did wonder if using the ai plugin to auto-flag such cases could also be a thing?
Just to note, and without wanting to take things off Topic:
it’s very hard (read near impossible) to detect AI contribution automatically. (Source: see the debate here)
in any case some people who’s first language isn’t the main language of the site (e.g. English) may use AI to rewrite their contribution, so even if detectable(?!) or “suspiciously well written”, would you really want to moderate or raise these for moderation?
why would anyone bother unless trying to Spam, in which case there is already a flag for that?
Had to read that three times to get what it meant.
I naturally think of the ai tag as topics related to ai but then realized that as a moderator you are much more keenly aware of what each tag means and that the ai tag is used specifically for the Discourse AI bot and not as I intended as AI related.
Noting this for others that may have fallen into the same line of thought.