PRINCIPLE / 01
AI suggestions separated from final moderator decisions
In development · Phase 3 roadmap
AI can suggest.
Humans stay accountable.
The planned review queue is designed around a firm separation: a model may surface a concern, while a moderator owns the decision and the member retains a comprehensible path to appeal.
PRINCIPLE / 02
A review queue designed around reasons, evidence, and appeals
PRINCIPLE / 03
Explicit retention controls for message content and AI verdicts
Honest status
AI review, appeals, content logging, and credit metering are planned work. No dates are promised, and these capabilities will remain outside launch copy until their engineering and legal gates close.