# AI powered Spam detection
**URL:** https://meta.discourse.org/t/ai-powered-spam-detection/343543
**Category:** Announcements
**Tags:** ai, spam
**Created:** [December 20, 2024, 4:56am UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543 "2024-12-20T04:56:23Z")
**Posts on this page:** 12
**Page:** 1
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 20, 2024, 4:56am UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/1 "2024-12-20T04:56:23Z")
I am happy to announce a new spam detection module in Discourse AI š
Over the past few months, we have sadly noticed that Akismetās performance has been quite uneven.
AI-based spam detection allows you to better control how spam scanning works in your community.
We have seen very good results on a few forums now. Please share your experience here.
For those self-hosting or generally looking for a āfreeā way of using these tools, have a look at [OpenRouter](https://meta.discourse.org/t/openrouter-support-for-discourse-ai/341531). There are quite a few free LLMs out there that can do a reasonable job scanning for spam.
Read all about it at:
> [@Discourse AI - Spam detection](https://meta.discourse.org/t/configuring-and-managing-ai-powered-spam-detection-in-discourse/343541):
>
> bookmark This guide explains how to configure and use Discourse AIās spam detection feature, including the setup process, scanning criteria, classification logic, customizations, and contrasts with AI triage. person_raising_hand Required user level: Administrator megaphone This is now default turned on for [Starter and Pro customers](https://meta.discourse.org/t/discourse-ai-spam-detection-replaces-akismet-plugin/354602), as well as our legacy Basic, Open Source, Creator, and Business customers. Discourse AI provides an efficient spam detection feature that identifies andā¦
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 22, 2024, 10:52pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/2 "2024-12-22T22:52:59Z")
Reporting back from results on meta over the weekend:
@Falco amended our ācustom instructionā here to:
> New users posting about crypto and linking to ANY site without adding important information to the conversation are spammers.
>
> Treat wildly off-topic posts as spam. As you are analyzing posts from what is primarily a support forum for Discourse Forum Software, if a topic isnāt about Discourse and related subject like other forum software, online communities, computer science and software development and doesnāt have a clear link to our main subject it is probably spam. For example, a new topic explaining details about a new car, a video game or any completely unrelated subject to Discourse, like vaccines, geo-politics or beach vacations are spam. Even technical topics, about subjects not related to Discourse are spam.
We have had **zero** missed spam over the past 2 days, we did catch:
* * *
We have seen so much slip through Akismet in the past few months, this new system is proving very effective.
We have **disabled** Akismet on meta.
---
### Author:  [@anurag](https://meta.discourse.org/u/anurag)
#### Post date: [December 29, 2024, 6:50pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/5 "2024-12-29T18:50:42Z")
I canāt enable this on a hosted Business account because even though the default LLMs included in the hosted version show up as āConfiguredā, theyāre not available to select on the āSpamā tab:
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 29, 2024, 9:47pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/6 "2024-12-29T21:47:47Z")
So sorry Anurag,
@awesomerobot and team working on improving clarity around the āseeded LLMsā.
We have been rolling out experimental features and āunlockingā the ability to use our internal LLMs when you opt-in for an experimental feature via the āWhatās newā feed in admin.
I completely hear you though that this is confusing
1. It should be clear that - LLM is only available for X,Y,Z features on the LLMs page.
2. We should invite you to āopt inā to an experiment direct from the LLMs page
If you head to your Whatās new feed (as a CDCK hosted customer) you will see:
Once you click that it will āunlockā the Small LLM for use with the anti-spam feature.
---
### Author:  [@Moin](https://meta.discourse.org/u/Moin)
#### Post date: [December 29, 2024, 10:04pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/7 "2024-12-29T22:04:39Z")
> [@sam](#):
>
> Once you click that it will āunlockā the Small LLM for use with the anti-spam feature.
At least for me, the āquick enableā feature for spam detection works quite confusingly. After a reload, it always appears disabled, regardless of whether the feature is actually enabled or disabled. I need to check the staff action logs to find out. The toggle always says āfeature enabledā after a reload, even when it was already enabled and will actually be disabled.
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [December 29, 2024, 11:16pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/8 "2024-12-29T23:16:36Z")
Thanks for reporting! We will have a look at this and report back.
---
### Author:  [@Saif](https://meta.discourse.org/u/Saif)
#### Post date: [December 30, 2024, 6:26pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/11 "2024-12-30T18:26:51Z")
@Moin weāve identified the issue, it was related to site setting naming. We have fixed it and are now deploying it for all sites soon
For now, If you go to `What's New` and click Check for updates and then go through the process, it should be working properly
---
### Author:  [@NateDhaliwal](https://meta.discourse.org/u/NateDhaliwal)
#### Post date: [December 31, 2024, 11:15am UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/12 "2024-12-31T11:15:09Z")
I must say, this spam detection is quite efficient! I tested it on my forum with a test user and it hid the post almost immediately.
Thanks, Discourse Team!
---
### Author:  [@per1234](https://meta.discourse.org/u/per1234)
#### Post date: [January 7, 2025, 11:21pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/14 "2025-01-07T23:21:54Z")
> [@anurag](#):
>
> even though the default LLMs included in the hosted version show up as āConfiguredā, theyāre not available to select on the āSpamā tab:
I also ran into this problem and couldnāt find a solution by searching for the āNo LLMs availableā key phrase, since it was previously only present in a screenshot, not in text form.
So Iām adding this reply to allow others who suffer from the same confusion to more easily find the solution provided here:
> [@sam](#):
>
> If you head to your Whatās new feed (as a CDCK hosted customer) you will see:
>
> 
>
> Once you click that it will āunlockā the Small LLM for use with the anti-spam feature.
---
### Author:  [@nickdb](https://meta.discourse.org/u/nickdb)
#### Post date: [January 10, 2025, 11:36am UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/15 "2025-01-10T11:36:22Z")
Trying to add custom instructions on our hosted site, which test fine, saving the changes gives an error āAn error occurred: You canāt use this model with this featureā
---
### Author:  [@sam](https://meta.discourse.org/u/sam)
#### Post date: [January 10, 2025, 7:42pm UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/16 "2025-01-10T19:42:37Z")
Apologies, I just queued a deploy which will fix it
Can you try again in 15 minutes?
---
### Author:  [@nickdb](https://meta.discourse.org/u/nickdb)
#### Post date: [January 11, 2025, 6:46am UTC](https://meta.discourse.org/t/ai-powered-spam-detection/343543/17 "2025-01-11T06:46:27Z")
Thanks, issue is resolved. Much appreciated