As the title says, i recently got banned from lemmy.world because they use regex (or a similiar mechanism) to ban someone instance-wide even if you write that thing on another instance
The mod log for whoever is interessed: https://feddit.it/modlog?page=1&actionType=ModBanFromCommunity&userId=20251941
Anyway, the culprit comment was this: https://feddit.it/comment/20981739
Apparently it banned me because i cited the blacklist that contained some URL or word that .world doesn’t like
Also, i had to use this 2nd account to write this post because i couldn’t access this community in my main one (it’s not an instance problem because when logged out i can find the community on feddit.it)
Honestly, i hope that .world admins wake up and understand that automated bans with regex are not the way and doing that on other instances is even worse!
Sadly that’s the issue with automoderators. Our own threativore has built-in functionality where you can PM it directly with the moderation-action code and the admins get a notification to reverse easily.
Honestly at this point it’s looking nothing like automod and entirely like a surreptitious attempt to control the damned fediverse
our team also gets notified about automod actions, but sometimes there is nobody available to deal with it immediately. we typically undo false-positives within an hour, and for most users it doesn’t have to be requested, as it’s obvious from context when it’s legitimate content.
Hm, if you have to review all auto-mod actions, I can imagine that at some point people will stop paying attention. I find it better to be notified only when you have to do something (such as with an appeal request)
we currently don’t have that many automod actions that it would be problematic. often we don’t even have a single one in a day. the notifications also include additional information about the user and reason to help with the review.