88
$\begingroup$

You may or may not have heard that there are some controversies around moderation on the network, specifically in relation to the moderation of GPT-produced posts, but more generally around lack of consultation with moderators (and emblematically, some policy changes which we shouldn't speak about publicly, in our understanding). There is a post here explaining the situation in more detail than I am able to.

A group of moderators from around the StackExchange network has signed an open letter calling for a moderation strike until the company is willing to address these concerns (EDIT: as of 9th June, 109 moderators out of the total pool of 538, from 76 different SE sites). I've signed the open letter and thus will be ceasing moderation while the strike continues (I've not received confirmation from all other MO moderators on their plans, but I know several of us will sign). We don't expect this to have a very noticeable negative impact on the site, but we wanted you to be aware.

If you agree with it, all users are welcome to sign the open letter. We will keep the community updated as events progress, and welcome your feedback here.

$\endgroup$
14
  • 18
    $\begingroup$ I have signed the letter. $\endgroup$
    – David Roberts Mod
    Jun 5, 2023 at 14:27
  • 18
    $\begingroup$ I've signed too. $\endgroup$
    – Todd Trimble Mod
    Jun 5, 2023 at 14:45
  • 7
    $\begingroup$ Let me note that "a group of moderators" sits, at time of commenting, at 82, together with a number of very high-profile former moderators (eg Shog9, who has 439k rep on the network-wide meta.SE, some may remember from days past). It also includes the people who run certain key pieces of software that helps keep the place tidy, and the software itself. $\endgroup$
    – David Roberts Mod
    Jun 5, 2023 at 14:49
  • 18
    $\begingroup$ in this thread from a few years ago meta.mathoverflow.net/q/4368/19860, the answer by Tim Campion suggests that MO should take some steps to assert its independence from SE. this seems like a good time to think more about such options $\endgroup$
    – ziggurism
    Jun 5, 2023 at 16:37
  • 7
    $\begingroup$ An inability to crack down on AI-generated nonsense would ruin MO, and I would second @ziggurism in suggesting that, if this policy from the higher ups at StackExchange is not reversed, MO should simply assert its independence. $\endgroup$ Jun 6, 2023 at 1:32
  • 1
    $\begingroup$ “…and emblematically, some policy changes our understanding is that we shouldn't speak about publicly.” Sorry, I’m not quite parsing this - do you mean there are policy changes you should not speak about publicly (this is what I suspect you mean), or that the policy change is that you shouldn’t speak publicly about policy changes in general? $\endgroup$ Jun 6, 2023 at 11:50
  • 3
    $\begingroup$ @Carl-FredrikNybergBrodda I mean there are policy changes coming down the road which I believe we aren't supposed to speak about publicly yet, and more generally there's quite a bit of frustration with SE leadership making public statements which don't really match what moderators have experienced or what they've told us. See, for example, here: meta.stackexchange.com/questions/389824/… $\endgroup$
    – Ben Webster Mod
    Jun 6, 2023 at 14:34
  • 6
    $\begingroup$ @SamHopkins I understand the impulse, but I do want you to think hard about the word "simply" there. It will not be simple at all. I don't think we're close to the point where we would decide to go our own way, but obviously, it remains a possibility should things really go downhill. $\endgroup$
    – Ben Webster Mod
    Jun 6, 2023 at 14:44
  • 3
    $\begingroup$ @BenWebster Does "asserting [MO's] independence" necessarily mean going our own way? Are MO's moderators bound by SE's policy decisions, or can MO assert independence with regard to policy while staying on the network? $\endgroup$ Jun 6, 2023 at 15:17
  • $\begingroup$ @YoavKallus That was what I understood the phrase to mean, but I agree that's not the only meaning it could have. Of course, we (both the moderators and high-rep users) are going to use our own judgement about moderation, which is a form of asserting our independence. But some things about the underlying function of the site, we cannot change without leaving the StackExchange network. $\endgroup$
    – Ben Webster Mod
    Jun 6, 2023 at 15:22
  • 6
    $\begingroup$ The linked answer by Tim Campion didn't call for leaving the StackExchange network, go our own way. Instead it suggested just to assert independence by writing a separate CoC and declaring it, not the SE CoC, to hold on MO. And to make a declaration about whether SE imposed mod firings. In the current context, we may also want to impose different rules about AI generated answers than SE. Those would be good steps to take, even if the more drastic step of leaving the network isn't necessarily going to happen. $\endgroup$
    – ziggurism
    Jun 6, 2023 at 15:59
  • 12
    $\begingroup$ This post by the mod team at Physics SE is an example of what it might mean to assert independence from SE oversight/control: physics.meta.stackexchange.com/questions/14438/… $\endgroup$ Jun 6, 2023 at 17:25
  • 6
    $\begingroup$ I've signed.... $\endgroup$
    – Tim Campion Mod
    Jun 8, 2023 at 12:47
  • $\begingroup$ I have Signed! :) $\endgroup$ Jun 17, 2023 at 1:31

3 Answers 3

59
$\begingroup$

It should be noted that users other than moderators can sign the open letter and likewise refrain from curation activities across the network, including MathOverflow, to the extent they feel comfortable, for instance not voting to close or delete questions, downvoting, editing etc etc. How or if anyone decides to participate is up to them, and they shouldn't be or feel pressured to join or not. Likewise, there are a number of Stack Exchange Inc staff who are caught in the middle, being used as mouthpieces here on the network by the company. Many of these are also unhappy with what is going on, and we should be careful to distinguish between frontline staff and SE Inc the company, in discussions. It is the policies and overall attitude towards moderators (being serially undervalued and poorly treated) that is largely the issue here.

The CEO has now gone on the record in the online media to openly claim things that are not accurate about how moderators across the board were treating claims of AI-generated spam. It is not the sign of healthy respect of a group of people who have contributed I estimate tens of thousands of hours of free expert labour, and who care about making the SE model work.

$\endgroup$
4
  • 2
    $\begingroup$ I certainly want to support you in the strike, so shall do as you suggest and not comment, vote etc. (I would have done this earlier if I had realised.) I notice that you don't recommend users abstain from asking or answering questions though, and this is also not specifically recommended in the FAQ at the end of the open letter. Just for my own interest, what is the thinking here? $\endgroup$
    – Nick Gill
    Jun 14, 2023 at 12:22
  • 1
    $\begingroup$ @NickGill - Curation of your own posts (questions and answers) is not moderation; it is content creation. Other systematic volunteer contributions to the quality of the site are viewed as community moderation, even if not performed with the involvement of an elected moderator. $\endgroup$ Jun 25, 2023 at 16:39
  • 1
    $\begingroup$ The ratio between the volume of content creation and the volume of moderation is supposed to determine the resulting site content quality somehow. Site quality can stabilize at anything between the dominance of neat and tidy computer generated babble or of insightful answers. That's the theory for some. $\endgroup$ Jun 25, 2023 at 16:39
  • $\begingroup$ Thanks @JirkaHanika, that's helpful. $\endgroup$
    – Nick Gill
    Jun 26, 2023 at 7:28
12
$\begingroup$

Stack Exchange have posted their analysis of the technical background about ChatGPT use etc here. In summary, they believe that there was a surge of ChatGPT-generated answers when ChatGPT first became available, but that it has now dropped to a very low level, and that a substantial number of users have been suspended when they either did not use ChatGPT at all, or used it only to improve their English. They base these conclusions on methods that are not reliable in individual cases, but which they believe to be reliable as a statistical aggregate.

It seems clear that Stack Exchange have handled their relationship with moderators and the community very badly in connection with this issue, and it is also possible that their analysis is wrong in various ways. Nonetheless, I do not believe that their analysis is stupid or frivolous, or that their interests are fundamentally misaligned with those of the community.

$\endgroup$
1
  • 8
    $\begingroup$ "but that it [the ChatGPT-answer surge] has now dropped to a very low level" this did not in fact happen. What happened was that the proxy used by the SE analysis dropped. The experience on the ground of moderators dealing with ChatGPT answers shows differently. The level is lower, but not the basically zero quantity they claim. From what moderators are saying, they don't disagree with the analysis if one keeps in mind the proxies being used, but they disagree the proxy is fit for purpose, especially as behaviour of users reacts to the pushback against ChatGPT answers. $\endgroup$
    – David Roberts Mod
    Jun 9, 2023 at 2:27
4
$\begingroup$

Ben Webster in the opening post: "If you agree with it, all users are welcome to sign the open letter. We will keep the community updated as events progress, and welcome your feedback here."

Am I correct to expect the keeping-the-community-updated shall happen here (in this thread)? Didn't see otherwise suggestive links anywhere...

[I signed the open letter as well.]

$\endgroup$
2

You must log in to answer this question.

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