Every claim should be judged by source strength. This page explains how we classify evidence across the site.
Source Strength Ranking
- Primary documents: court records, regulatory filings, research papers, company status pages, official statements, and original datasets.
- Named institutional reporting: articles from recognized publications with clear authorship and editorial review.
- Expert commentary: named researchers, attorneys, clinicians, engineers, or policy specialists speaking within their expertise.
- User testimony: direct experiences from users, forums, Reddit, emails, or submissions. Useful for pattern detection but not treated as independently verified fact by itself.
- Editorial inference: conclusions drawn from multiple sources. These should be signaled as analysis.
User Stories
User stories are valuable because they show patterns and lived experience, but they can be incomplete, emotional, or disputed. We preserve them as accounts and avoid converting them into proven institutional facts unless independent documentation supports them.
Updates
AI products, lawsuits, and safety policies change quickly. Pages should be updated when new filings, research, status reports, or company statements materially change the record.