A forum administrator, justin_gordon, is seeking solutions to manage a significant influx of spam on their forum. They have been unable to manually address each spam message due to the volume. pfaffman suggests using administrative bulk operations to delete the spam and questions why the spam bypassed Akismet. ondrej confirms there is no current way to bulk process flags but suggests granting a trusted member moderator privileges to help. justin_gordon also needs to bulk approve 590 messages. justin_gordon inquires about the Discourse AI Post Classifier plugin and its effectiveness compared to Akismet. ondrej explains that the plugin requires both the Discourse AI and Discourse Automation plugins and a Large Language Model (LLM), which can be expensive. pfaffman mentions that Akismet can be configured to apply to specific trust levels. rlauriston requests a feature for bulk processing of spam in the review queue. pfaffman provides a link to a solution involving AI review and deletion of spam posts. rlauriston expresses concern about automatic deletion by AI and the overwhelming number of spam messages in the review queue.
My forum Coding Forum - Developers Discussions and Problem-Solving on ShakaCode Forum got overrun with spam in the past few weeks!
It’s so much that I can’t manually go into every message.
I’m using
Should I install this:
Those all got past akismet? That seems surprising. Did they do something to get to tl1 and bypass akismet?
Since they are so similar you can look at Administrative Bulk Operations to see ways to delete them all (I think it should have an example).
That’s a lot of spam! As far as I know, there isn’t a way to bulk process flags. There is on open Contribute > Feature request though. I see you have 2 inactive admins on your site. I didn’t look at how active your community is, but have you considered granting a trusted member moderator to help you clear stuff up? I can’t see any ways to bulk process/delete flags in the Administrative Bulk Operations topic.
Not sure if this has changed since April but it would be a neat feature 
There is also this option?
What about that plugin? Akismet good enough?
What is that?
You need both the Discourse AI and Discourse Automation plugin for the post classifier to work. Additionally, you will need a LLM [1]. These can get expensive especially if it has to make a lot of calls.
I’d say Akismet is good enough at accurately detecting spam so its really up to you what you choose.
tl1 means Trust Level 1 which is a Basic user. See Understanding Discourse Trust Levels for more info.
Large Language Model ↩︎
And if you’re using akismet, it has a setting to determine which trust levels it applies to.
Is there a feature request somewhere I can vote for?
My site is getting slammed by what seems like Discourse-aware AI-generated spam. I’d like to be able to display the review queue with checkboxes, click Select All to select all of them, skim to uncheck any legitimate posts, and take the same action (delete and block) against them all.
This would be a good place to have a checkbox. Anyway a similarly concise list.
I describe how to get the AI to test it and then delete the ones it caught here: Spam: Retroactively get AI to review recent posts and then delete the ones marked as spam
It has the advantage of being a solution that’s available now. It’s far more complicated than a few clicks, though.
I’m dubious of having AI delete automatically as there might be a legit post somewhere among the messages.
Currently almost all the spam is going to the review queue, which I’m ignoring as it’s too much work to mark them spam one by one. Since yesterday, 210 messages have been added by the AI spam filter.
