Trust & safety
Trust and safety
Safety is structural: designed workflows, audit logs, and public accountability — not engagement-maximizing moderation.
Community guidelines
Participate in good faith. No harassment, coordinated manipulation, or impersonation. Criticize ideas with evidence. Respect deliberation norms and facilitator guidance.
Moderation process
Reports are triaged by trained moderators with appeals. Actions are logged with reasons. Severe or repeated violations may lead to restrictions after review.
AI red lines
AI must not vote, rank citizens, hide lawful opinions, or auto-publish submissions. Human review is required for consequential moderation and dossier finalization.
Transparency reporting
Quarterly aggregates of moderation actions, verification decisions, and safety incidents are published here. Demo environment shows sample aggregates only.