The 5 Pillars of Evaluation
Every submission, revision, and featured entry is assessed against these non-negotiable criteria. Content that fails to meet baseline thresholds in any pillar is flagged for revision or archival.
Accuracy & Verification
Claims must be traceable to primary or authoritative secondary sources. Our AI cross-references data in real-time, and human experts validate statistical, historical, and scientific assertions.
- Primary source citation required for all factual claims
- Statistical data verified against official databases
- Peer-reviewed journals prioritized for scientific content
Neutrality & Balance
Articles must present multiple perspectives on debated topics without endorsing a single viewpoint. Language is continuously audited for implicit bias or loaded terminology.
- Proportional representation of credible viewpoints
- Removal of promotional or advocacy language
- Contextual framing for historically sensitive subjects
Timeliness & Currency
Knowledge decays. Articles are assigned a freshness index based on topic volatility. High-impact fields (technology, medicine, policy) trigger automated review cycles.
- Quarterly reviews for volatile disciplines
- Version control with visible edit histories
- Deprecation warnings for outdated frameworks
Structural Integrity
Readability and logical flow are quantified. Articles must follow standardized taxonomies, include clear headings, and maintain appropriate depth relative to the subject complexity.
- Consistent citation formatting (APA/Chicago)
- Mandatory infoboxes, tables of contents, and glossaries
- Algorithmic readability scoring (target: 85+)
Cultural & Linguistic Fidelity
Translations and localized entries are reviewed by native-speaking subject experts to preserve nuance, avoid etnocentric framing, and respect regional terminology.
- Native expert validation for all translations
- Region-specific data normalization
- Indigenous knowledge attribution protocols
Provenance & Transparency
Every article displays its evaluation score, review timestamps, and contributor credentials. Hidden editing is prohibited; all changes are immutable and publicly auditable.
- Public contributor verification badges
- Blockchain-backed revision logs
- Clear disclosure of AI-assisted drafting
The Evaluation Pipeline
From initial submission to publication, every article passes through a hybrid human-AI review system designed to eliminate error while preserving editorial independence.
AI Pre-Screening & Structuring
Submissions are parsed for formatting compliance, citation completeness, and basic factual consistency. Plagiarism checks and tone analysis run automatically.
Domain Expert Assignment
Qualified reviewers matching the article's taxonomy are notified. Experts hold verified academic or professional credentials in the relevant field.
Peer Review & Cross-Reference
Two independent experts evaluate accuracy, neutrality, and depth. Discrepancies trigger a third-party arbitration review. Edits are tracked in real-time.
Quality Scoring & Publication
A composite score is generated. Articles scoring ≥ 8.5/10 are published with a "Verified" badge. Lower scores enter revision queues or archival.
Quality Metrics & Tiers
We don't use vague labels. Every article receives a quantitative quality score based on weighted evaluation criteria. Tiers determine visibility, featured placement, and API access priority.
Evaluation Criteria FAQ
AI handles initial syntax checks, citation formatting, plagiarism detection, and preliminary fact-crossreferencing. All substantive judgments, tone assessments, and publication approvals require verified human experts. AI suggestions are always visible and rejectable by reviewers.
Articles scoring below 7.0/10 are returned to the author with detailed, line-by-line feedback. Chronic failures or policy violations result in account review. Archival is reserved for historically preserved but scientifically outdated content, never for quality failures.
Yes. Our appeals process allows structured challenges within 30 days of publication. Appeals are routed to an independent editorial board with rotating membership. All appeal outcomes and reasoning are published transparently.
All contributors and reviewers must disclose institutional affiliations, funding sources, and potential conflicts. AI flags potential bias patterns, and the system automatically reassigns reviews when overlap is detected. Undisclosed conflicts result in immediate credential suspension.