The review workflow is what keeps the graph from becoming a rumor engine. A claim can be interesting, popular, or emotionally powerful and still not be ready for public framing.
What to notice
Statuses protect readersProposed, reviewed, contested, stale, corrected, rejected, retracted, sensitive, and published should be explicit.
Human review remains centralAutomation can suggest; editors decide what is safe, clear, and fair to publish.
Memory mattersCorrections and failed claims should remain visible enough to prevent endless recycling.
Discussion questions
- What should trigger a re-review?
- How can correction be visible without becoming punitive?
