2
$\begingroup$

My question here is about a particular post, but I think this may be of interest in other similar cases.

The post in question was titled Recurrence with square roots, and it was deleted by Community Bot, which is "a background process that helps keep this site clean!"

I have been unable to find specific reasons for this deletion. The earlier comment by Sam Hopkins may suggest that the reason might have been "an AI poster".

If so, I am wondering how this could have been detected in such a case -- by humans and by the (automatic?) background process. I would not be able to do this in such a case. Can anyone shed some light on this matter?

I am also wondering if it is possible to find specific reasons for Community Bot's specific actions, and if it is possible to challenge its actions.

$\endgroup$
10
  • 3
    $\begingroup$ As shown in the timeline, the post was deleted because the user account have been destroyed. This must have been done by a human, not by an automated process. All kinds of things are formally attributed to the Community user in the database. $\endgroup$ Commented Oct 7 at 16:49
  • $\begingroup$ @EmilJeřábek : Thank you for your comment. How can I find the timeline you are referring to? Also, I am still wondering how AI could have been detected in such a case -- by humans or any other means? $\endgroup$ Commented Oct 7 at 16:57
  • 3
    $\begingroup$ The clock icon under the voting buttons. A direct link is mathoverflow.net/posts/480183/timeline . $\endgroup$ Commented Oct 7 at 17:00
  • $\begingroup$ @EmilJeřábek : Thank you for your further response. $\endgroup$ Commented Oct 7 at 17:18
  • 5
    $\begingroup$ While I am not the moderator who destroyed this particular account, what I can tell is that often there are certain patterns in the data which is only visible to moderators which make quite clear that an account belongs to a series of accounts created to spam the site with AI-generated posts. $\endgroup$
    – Stefan Kohl Mod
    Commented Oct 7 at 18:01
  • $\begingroup$ @StefanKohl : Thank you for this information. However, I think it would be useful for regular MO users to have some tools/knowledge to identify AI production, say so that not to waste their time answering AI-generated questions. In this particular case, I am still unable to recognize the post in question as AI-generated, whereas at least one user (Sam Hopkins), who is not a moderator(?), was able to have such a suspicion. So, I am still wondering how/if regular MO users can detect AI in such cases. $\endgroup$ Commented Oct 7 at 18:45
  • 3
    $\begingroup$ I presume the AI flag was raised in response to a set of questions posted by this unregistered user, rather than in response to a single question; to fight this, we need meta.mathoverflow.net/questions/6042/… $\endgroup$ Commented Oct 7 at 19:05
  • 4
    $\begingroup$ @IosifPinelis What regular users can use as criteria is e.g.: AI spammers are usually new accounts, very often unregistered; often they post a number of questions or answers within a short time, sometimes about quite different topics. Sometimes at least one of the posts made by the respective account looks "AI-like" in terms of style, formatting or particular formulations, or it contains mathematical mistakes a human mathematician hardly would make. $\endgroup$
    – Stefan Kohl Mod
    Commented Oct 7 at 19:39
  • $\begingroup$ @CarloBeenakker : Thank you for your comment. $\endgroup$ Commented Oct 7 at 21:00
  • $\begingroup$ @StefanKohl : Thank you for your further response. $\endgroup$ Commented Oct 7 at 21:00

0

You must log in to answer this question.

Browse other questions tagged .