How do you manage flags on non-human posts?

AI-generated summary

In the Community Building category, Moin initiates a discussion regarding the management of flags on non-human posts, such as those from system accounts, Discobot, or AI bots. Moin highlights that these flags do not appear in the standard review queue, creating a risk that they may be overlooked. This oversight can lead to frustration for users who report issues and potentially result in the deletion of entire topics if system posts are flagged as spam and hidden without moderator intervention. Moin also notes that some system posts originate from deleted users and that AI bot PMs are often encouraged to be reported for inaccuracies. The user asks for best practices, such as using Data Explorer queries or checking the moderator inbox, to ensure these flags are not missed.

hugh responds to the initial post, acknowledging the concern and stating that flagged system posts should ideally be added to the review queue like any other. They commit to looking into making this change. However, a year later, Moin revisits the topic after flagging a bot post for testing. Moin observes that while a message appeared in the moderators’ inbox, there was still no entry in the review queue. The user reiterates their curiosity about how others handle this situation and asks if there are alternatives to Data Explorer queries for finding flags on system or Discobot posts that do not trigger a private message to moderators.

Flags on non-human posts such as those from system, Discobot, or AI bots do not appear in the review queue. I am interested in how you handle them.

Primarily, I’m curious about how you ensure they aren’t missed. After all, it’s frustrating for users if they report something and it seems like nobody cares.

Do you have a Data Explorer query to find such flags? Do you regularly check the moderator group’s inbox? If you rely on the latter, how do you also ensure that a flagged post is not hidden or deleted somewhere because of, for example, too many spam flags?

Of course, one might say that system posts shouldn’t be spammy, but a user could still cause damage by flagging them, which can lead to them being hidden and, if no staff member intervenes, deleted. And since those flags do not appear in the review queue, there is a real risk they will be overlooked and entire topics disappear.

In addition, some posts authored by system actually originate from deleted users, which may later turn out to contain spam.
And then there are PMs with the AI bots. Users are often encouraged to report these if they contain incorrect or misleading information.

Here on Meta, also some documentation topics are posted by system.

So, how do you ensure that these kinds of flags aren’t overlooked?

7 Likes

Thanks for flagging this, @Moin.

IMO, flagged system posts should be added to the review queue like any others, so I think that’s a change worth making. We’ll look into this!

3 Likes

I would still be interested in how you have handled it so far. For example, on ask.discourse.com you also ask us to report mistakes.

A year has passed.

I just flagged a bot post again for testing purposes and then wondered why there was a message in the moderators’ inbox but no entry in the review queue.

I’m still curious to know how others handle this. Is there an alternative to a data explorer query to find flags on posts by system, discobot, … that don’t send a PM to moderators?

2 Likes