35

Currently, if a question or answer accumulates a sufficient number of offensive or spam flags against it, it is automatically deleted and a reputation penalty is imposed on the poster. Likewise, comments that have a sufficient number of flags against them are auto-deleted. This requires no moderator intervention, and if no moderator happens to be active when this occurs, we won't even know that this happened.

These are cases that moderators should be made aware of. If someone posted spam obvious enough to be flagged heavily by the community, odds are that account (and any related account) needs to be looked into. Trolling or abusive posts and comments are cause for a strong warning from a moderator, if not a suspension.

And then there's coordinated gaming of these flags to attack others. The use of sock puppets to cast synchronized flags in an attempt to destroy content once was a purely theoretical exploit. That is no longer the case. In the last year, I've personally dealt with two users who employed sock puppets to flag and auto-delete comments critical of their answers, and both attempted to used offensive flags to destroy competing answers (one succeeded). Over the last couple of days, GameDev has been dealing with a much nastier form of this.

For these reasons, I would suggest that moderators be made aware of comments and posts automatically deleted by the community as a result of spam or offensive flags. For posts, this could come in the form of a flag raised by the Community account similar to the warnings we get about users mass-deleting posts, etc. Something like "Post deleted due to community spam / offensive flags" would make it clear what happened here. Warning flags would not need to be generated if a moderator was the one to cast the auto-deleting spam or offensive flag.

For comments, I'm not sure about the best way to expose this, whether by a Community flag on the post ("XX offensive comments deleted") or by a special type of comment flag indication in the mod dashboard. Comment notifications may also need to ignore ones that get flagged and auto-deleted based on containing certain non-offensive keywords (the "what have you tried" type comments).

There have been a number of spam, trolling, or otherwise abusive incidents over the last year that could have been dealt with much earlier if we had this kind of information presented to us. I do not believe that it would be any noisier than the other automatically raised warning flags we get now.

Improve this question
5
  • 4
    See also we need better spam detection tools, in which I propose a parallel feature ("A listing in the moderator tools of posts recently marked as spam/offensive").
    – nhinkle
    Jun 18, 2013 at 16:23
  • 2
    Given a site like Stack Overflow, there are few things I detest more than means of abuse being dismissed as theoretical. Any scenario I conjure at this moment can and will be executed by someone eventually - and in that kind of situation, prevention is always better than cure.
    – BoltClock's a Unicorn
    Jun 18, 2013 at 16:36
  • 1
    I wonder if it might be a good idea to show such in the 10K tools. If 6 (or whatever) 10K users say "deletion is pukka", and none says "dubious", odds are, the deletion is indeed correct. Of course for that to have a chance of diminishing the work for moderators, such incidents would have to be withheld from direct moderator notification for 1, 2(?) hour(s) [unless say 3 "dubious" votes are cast before], which would be bad if an incident required urgent moderator action.
    – Daniel Fischer
    Jun 18, 2013 at 17:46
  • @BoltClock'saUnicorn "All attacks are theoretical until they're practical."
    – voretaq7
    Jun 18, 2013 at 20:28
  • They ARE for "in need of moderator attention" flags
    – John Militer
    Feb 24, 2017 at 20:47

2 Answers 2

Reset to default
8
+50

The moderator dashboard -- what you get to if you click the "mod" link in the topbar -- shows information about all moderators and any SE employees -- every user with a diamond -- who took moderation actions during the relevant time period. From each of those entries we can see history that includes flags handled.

How about a line in that chart for the Community user, another user with a diamond? This would allow mods to review Community-resolved spam flags (and some other things), so we could take any addition action that's needed.

We (now) get auto-flags for users who are the recipients of too many abusive/not-constructive flags on comments. This should catch cases where a group of users is targeting some other user, like the person they're arguing against in a comment thread. So I think if we can also have a way to look into Community's flagging activity to investigate other cases that come up, that should give us what we need.

Improve this answer
6

I agree with the idea of something on the mod tools links page -- A "Recent actions by Community" page is something that could be useful to look at on an as-needed basis.

I'm less enamored with the idea of presenting this info as a flag or other "active" alert - If we do that it seems to defeat the purpose of the "community action" mechanism (because mods would still have to look at everything).

That could be mitigated by "special case" coding so it only shows if N posts/comments by someone have been deleted, where N is >2 or something, but the value of N would be site-specific and annoying to tweak.

Improve this answer
2
  • 1
    I see the advantage of the auto-deletion being that the community can act quickly to remove really problematic content, not that it lessens the moderator workload. It's trivial for us to delete spam or obviously offensive content, but the auto-deletion prevents troublesome content from being a distraction if we're not around. We can quickly dismiss these flags if there's nothing more to do after the community has taken action. I don't see this adding much to our load.
    – Brad Larson
    Jun 18, 2013 at 17:56
  • 1
    @BradLarson My main concern is not timeliness, it's noise. Rule #1 of monitoring systems is "Only alert on things that require action", the same applies to mod notification: If the alerts are noise they're worse than useless, they habituate us to ignore them ("because it's always fine"). If it's just "made available" as opposed to an active alert like a flag we sidestep the whole issue nicely.
    – voretaq7
    Jun 18, 2013 at 19:04

You must log in to answer this question.

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