Moderation
AI moderation, rules of engagement, and what happens when a message is flagged.
Why there is an AI moderator
Every Dialog Lab conversation includes an AI moderator that stays in the chat and keeps things civil. It steps in when a message crosses the line — but vigorous disagreement about ideas is encouraged.
As the app explains: “Time-limited, anonymous. An AI moderator stays in the chat and keeps things civil — it’ll step in if a message crosses the line.”
Meet your moderator
Effective moderation takes both trust and skill. Our AI moderator:
- Never takes a side — it referees how you argue, not what you believe
- Blocks abusive messages before they're ever sent, so they never touch the conversation
- When it steps in (a nudge, a flag, a suggested rephrase), you both see it and why
- Makes sure you both get your turn
Rules of Engagement
During onboarding, you agree to four simple ground rules. Keep these in mind throughout every conversation:
Ground Rules
- Assume good intent in your partner
- Respect the conversation style (discussion or debate)
- Labeling or stereotyping will be flagged
- Slurs and abuse will be blocked
Warn vs block
The moderator evaluates every message before it is delivered. Two outcomes are possible:
Warning
The message is sent. A neutral reminder appears to both participants — for example, nudging you back to the claims rather than motives. Warnings are shared; neither person is singled out.
Block
The message is withheld and not delivered. Only the sender sees the blocked message and a private explanation of why it was stopped. The chat shows: “Blocked message from sending.”
| Allowed | Not allowed |
|---|---|
| Strong disagreement on ideas | Personal attacks and insults |
| Challenging your partner’s reasoning | Slurs and dehumanizing language |
| Sharing personal experiences | Harassment or threats |
What you see in chat
Moderator messages appear labeled AI Moderator. Some notes are marked Only visible to you — these are private coaching or block explanations sent only to the sender.
The moderator may also remind you of ground rules at the start of a conversation. For longer messages, it may suggest a clearer rephrase so your partner can follow your point.
Pause anytime
Either participant can send a Pause from the chat menu to let the other person know they do not expect to reply for a while. It is a courtesy heads-up — not a freeze on the conversation. Either of you can keep chatting if you want to.
Reporting vs auto-moderation
The AI moderator handles message-level civility in real time. If your partner’s behavior is concerning beyond individual messages — harassment, bad faith, or repeated civility violations — use Report Conversation from the chat menu or post-chat review. That sends a report to the Dialog Lab team for human review.
See How to Get Support for reporting details.