# Experiments with AI based moderation on Discourse Meta

**URL:** https://meta.discourse.org/t/experiments-with-ai-based-moderation-on-discourse-meta/357865
**Category:** Community Building
**Tags:** moderation, ai
**Created:** [March 19, 2025, 12:31am UTC](https://meta.discourse.org/t/experiments-with-ai-based-moderation-on-discourse-meta/357865 "2025-03-19T00:31:54Z")
**Posts on this page:** 1
**Showing post:** 10

<div class="post-metadata">

### Author: ![RGJ](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/rgj/32/523185_2.png) [@RGJ](https://meta.discourse.org/u/RGJ)
#### Post date: [March 25, 2025, 12:30pm UTC](https://meta.discourse.org/t/experiments-with-ai-based-moderation-on-discourse-meta/357865/10 "2025-03-25T12:30:34Z")

</div>

> [@Ed\_S](#):
>
> Just an idea, could you do something in your prompt so the moderation bot will at least once in a while post a ping response, to show it’s working. Maybe, for example, with 1% probability when a post needs no action, post a note that this post needed no action. Or a lower probability, for a busier forum.

Alternatively, have it trigger on an innocent, common-but-not-too-common word. “Flag posts that mention pineapples”.

---

_[View the full topic](https://meta.discourse.org/t/experiments-with-ai-based-moderation-on-discourse-meta/357865)._
