Moderation¶
Approval before a community quiz can be shared with anyone.
Decision: ⚪ idea
Context¶
The author can always take their own quiz. Sharing it requires our approval - the gate is enforced by the authoring flow state machine, not by convention.
Review checks:
- no illegal content, hate speech, or targeting of private individuals
- questions are answerable and not push-polled
- orientations and weights are coherent enough to produce a meaningful result
- result modules are configured and the quiz actually finishes
- not spam, not a near-duplicate of an existing quiz
Outcomes:
- approve -> quiz published
- reject with a written reason -> quiz need to be fixed
- reject and remove -> quiz is outside acceptable boundaries so we need to remove it
Opportunity¶
- Everything published under our name is our credibility, and credibility is the whole asset - see Why. A gate is what makes opening the editor survivable.
- Provides a quality floor in the window before quality score has enough data to rank anything.
- Written reasons double as creator education - the cheapest way to raise the median submission.
Risk¶
- Manual review is a human bottleneck on a volunteer team - the exact constraint this module exists to escape.
- Judgement calls on political content will be read as bias, so criteria must be published and reasons written down.
- Volume from LLM authoring can outpace reviewers faster than we can staff them.
- Latency demotivates creators at their most motivated moment.