> For the complete documentation index, see [llms.txt](https://help.steeple.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.steeple.com/en/news/news-moderation-ia.md).

# Automatically moderate posts and comments with AI

When moderation is enabled, Steeple automatically checks every post and comment before it is published. The content is analyzed by AI according to rules you can customize. If it doesn’t comply with them, it isn’t published.

### Enable or disable moderation

Moderation is **enabled by default**, for both posts and comments.

1. Go to **Posts** > **Settings** > **Moderation**.
2. Enable or disable the options:
   * **Post moderation** : AI checks each post before it is submitted.
   * **Comment moderation** : AI checks each comment before it is submitted.
3. Click **Save**.

The two options are independent: for example, you can moderate comments only.

<img src="https://608007981-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FByS0e17eM732OXdGGkme%2Fuploads%2F9kiirrY9UUxko42xtPZN%2Fimage.png?alt=media&amp;token=203318cc-7359-4370-be4d-684d883a5ac9" alt="" data-size="original">

### Default rules

AI blocks content that contains:

| Rule                            | What’s prohibited                                                                                   |
| ------------------------------- | --------------------------------------------------------------------------------------------------- |
| Insults and profanity           | Insults, aggressive profanity, and degrading language                                               |
| Harassment and personal attacks | Targeting, belittling, or ridiculing a person or group                                              |
| Discrimination                  | Statements relating to origin, gender, religion, sexual orientation, age, disability, or appearance |
| Explicit sexual content         | Explicit statements or inappropriate sexual innuendo                                                |
| Threats and violence            | Threats, incitement to violence, or glorification of violent acts                                   |
| Sensitive personal information  | Another person’s health, salary, address, or private life, without their consent                    |

These rules apply to both posts and comments. AI checks the text (title, description, and poll responses), not images or files.

### Customize the rules

The rules are written in plain language: AI reads them as instructions. You can adapt them to your organization, up to a limit of 5,000 characters.

* **Edit or delete a rule** : edit the text directly in the **Moderation rules**.
* **Add a rule** : write it on a new line, using the same format. For example: “- Competitors: don’t mention our competitors by name.”
* **Restore the original rules** : click **Reset**.

Remember to click **Save**.

Tip: AI understands a short, precise rule with a concrete example better than a vague one.

### What happens when content is blocked?

* The author sees a message explaining that their content doesn’t comply with the moderation rules, and why.
* They can edit their text and submit it again.
* Administrators are not notified.

### Frequently asked questions

**Can AI make mistakes?**

Yes, like any automated system. If legitimate content is blocked, rephrase it or adjust your rules.

**Is previously published content checked?**

No. Moderation applies to new content and edits when they are submitted.

**Who can change these settings?**

Only organization administrators.

**Can I adapt the rules to specific cases?**

Yes. For example, the default rule prohibits sharing another person's address or private information without their consent; if your colleagues share their work phone numbers with each other, specify in the rules that this is allowed.
