> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.dumgum.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Fictional Disclosure

Configure personas to disclose they are fictional profiles when a user questions their authenticity, for legal compliance and platform transparency.

## Overview

Some jurisdictions legally require platforms to disclose that an account is fictional when a user sincerely asks. Fictional Disclosure lets you meet that requirement without breaking the conversation entirely.

When enabled, the persona responds to authenticity probes with a brief, controlled disclosure and an invitation to keep chatting. If the user continues to probe, the standard AI Suspicion ban is applied.

***

## How it works

Fictional Disclosure is built on top of the **AI Suspicion guardrail**, which detects when a user is probing the authenticity of the persona (e.g. "are you real?", "are you a bot?", "you're not a human, are you?").

When both are active:

1. The AI Suspicion guardrail detects the probe
2. The persona responds with a disclosure (see [What the persona says](#what-the-persona-says))
3. If the user accepts the disclosure and moves on, the conversation continues normally
4. If the user keeps probing, the standard AI Suspicion ban is applied and the conversation is closed

```
User probe detected
        ↓
fictionalDisclosure ON?
        ↓ yes
Persona discloses → conversation continues
                  → user probes again → ban applied
```

<Note>
  If the **AI Suspicion guardrail is disabled** on your project, Fictional Disclosure will not trigger: no disclosure and no ban will be applied. Both must be active for the feature to work.
</Note>

***

## What the persona says

The persona uses only one of the following phrases:

> **"This is a fictional account"**
> **"This is a fictional profile"**

It will not expand on this. Follow-up questions about what is real or fake are redirected warmly back to the conversation.

**The persona will never say:**

* "I am not a real person"
* "I am not human"
* "I am an AI / bot"
* Any equivalent in any language

***

## Dependencies

| Requirement                | Detail                                     |
| -------------------------- | ------------------------------------------ |
| AI Suspicion guardrail     | Must be **enabled** on the project         |
| `fictionalDisclosure` flag | Must be set to `true` in `replyParameters` |

Fictional Disclosure has no effect if AI Suspicion is turned off.

***

## Enabling the feature

Set the `fictionalDisclosure` flag in `replyParameters`:

```json theme={null}
{
  "replyParameters": {
    "fictionalDisclosure": true
  }
}
```

The AI Suspicion guardrail must also be enabled (it is by default).

***

## What happens after disclosure

After disclosure, if the user accepts and moves on, the conversation continues normally. If the user keeps probing, the standard AI Suspicion ban is applied.

<Note>
  The persona's meeting refusal rules remain active regardless of disclosure.
</Note>

***

## Use case

This feature is designed for legal compliance in jurisdictions that require platforms to disclose the fictional nature of a profile upon request. It allows the platform to fulfil that obligation in a controlled, consistent way, without exposing infrastructure details or breaking the platform experience more than necessary.