43

This a follow-up on Sites can now request to enable a banner to warn about their policy on AI-generated content where I posted a now-deleted answer to comply with the guidelines. We frequently see users posting nonsensical answers here on Meta Stack Exchange, clearly generated by an AI which not having a clue what it's talking about. I'd gladly see the first of the two proposed banners:

Reminder: Answers generated by Artificial Intelligence tools are not allowed on Meta Stack Exchange.

here. The announcement says:

If this is something you think your community is interested in opting into, please start a discussion in its Meta site to get community consensus on the appropriate course of action

Since Meta Stack Exchange is its own site, I'm throwing this question as a poll (upvote = I agree, downvote = I disagree - hence the tag). Additional Meta Stack Exchange-specific considerations can be discussed in comments / answers.

6
  • 4
    Out of curiosity, how often does AI-generated content get posted on MSE currently?
    – V2Blast
    Jan 5 at 20:33
  • Would it be dismissible?
    – W.O.
    Jan 5 at 22:17
  • 1
    @W.O. yes according to meta.stackoverflow.com/q/426353/395857. Better be. Jan 5 at 22:40
  • 6
    @V2Blast anecdotally, somewhere between 5-6 a week at its peak, and maybe 1-2 now. I figure these folks care little enough about meta to try their luck. This dosen't include the much more numerous spammers who use it to create filler for the mechanically seperated meat product. Jan 6 at 5:46
  • Alternatively, we could not, and let these users post obvious AI answers here, which we then use against them when they post AI answers on other sites…
    – Laurel
    Jan 6 at 13:18
  • As I see it - if they ignore it, we get that anyway, and if they don't it means they actually read and it works. Feels win win to me. Jan 7 at 0:58

2 Answers 2

18

Yes please.

Both as a former moderator, and as an active user, there's never been any redeeming value to GenAI meta posts here.

Generally they're vapid, generic non-answers, and while I am not confident that the sort of folks who think posting stuff right off the prompt is fine would care enough to read the warning label, we should at least make it clear that it's not welcome on this site.

4

Given the particular focus of meta sites (on the network's policies rather than on knowledge more broadly), I struggle to imagine how AI questions or answers could ever be useful on Meta—either as a whole, or on site-specific metas. Would whatever decision we make here be worth applying to site-specific metas as well?

I haven't seen any AI-generated meta posts on the sites I moderate, but they're also on the smaller side, so I don't know if this is a problem for the bigger ones.

4
  • 8
    It's network policy that enabling/disabling the notification that's being discussed here is a decision that must be made on a per-site basis. Particularly given that if AI-generated content is permitted at all on a site is a per-site policy, it's inappropriate for us to make the choice here on MSE for the notification to be enabled on all per-site metas. While I would certainly prefer that SE had taken the stance of just not permitting AI-generated content across the network, we shouldn't try to impose this notification on all meta sites under the current conditions.
    – Makyen
    Jan 6 at 1:00
  • @Makyen I respectfully disagree. While I do support decisions being made on a site-by-site basis in the majority of cases, AI-generated content is not only a very clear exception, but it is one of those that stick out far more clearly than the rest.
    – 000
    Jan 6 at 10:49
  • 1
    @ElEctric I understand your disagreement, but allowing each site to make their own decision with respect to allowing AI-generated content or not is the policy the company has adopted. This is not really the question and answer on which to be expressing disagreement with that. Given that staff at the company are the ones who will need to turn on these notifications, and they won't turn them on network wide based on an answer here, it doesn't do any good to attempt to do an end-run around the established policy by trying to impose these notifications where the site hasn't chosen for themselves.
    – Makyen
    Jan 6 at 15:49
  • @Makyen I understand that sites might make different decisions, but meta seems like a clear exception to me—what's appropriate on Physics.meta seems like it should align with what's appropriate on Meta, not what's appropriate on Physics. (ChatGPT might seem to know something about physics, but not about Physics.SE policy.)
    – Draconis
    Jan 6 at 17:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .