Aevum Encyclopedia Transparency Report
This report details Aevum Encyclopedia's commitments to editorial independence, data privacy, AI transparency, content moderation, and legal compliance. We publish this annually to maintain trust with our 180,000+ contributors, researchers, and global readership.
Executive Summary
Aevum Encyclopedia operates on a foundation of verifiable knowledge, algorithmic transparency, and editorial autonomy. In 2024, we processed over 2.4 million active articles across 140+ languages, served 412 million monthly readers, and maintained a 99.94% content accuracy rate according to third-party academic audits.
Our AI systems assist with cross-referencing, semantic search, and draft flagging, but all published content undergoes human expert review. We received 3,842 legal requests, complied with 3,105 (80.8%), and published full details where legally permissible. We also expanded our independent ethics board and open-sourced our content moderation policy framework.
💡 Key Commitment: We will never edit, suppress, or alter verified content at the direction of any government, corporation, or third party without judicial oversight and public disclosure.
Legal & Government Data Requests
We receive requests from governments, law enforcement, and courts for user data, content removal, or metadata. We review each request against our Legal Compliance Framework and challenge overbroad or vague demands.
| Category | Requests Received | Compliance Rate | Challenged / Denied | Primary Jurisdictions |
|---|---|---|---|---|
| User Account Data | 1,420 | 82% | 256 | EU, US, UK, Canada |
| Content Removal | 1,185 | 68% | 379 | Asia-Pacific, MENA |
| Server / Metadata | 892 | 91% | 81 | North America, EU |
| Copyright / IP | 345 | 94% | 21 | Global | d>
All compliant requests were processed within 72 hours. We published 1,247 transparency notices across affected articles when legally allowed.
Content Moderation & Enforcement
Our moderation system prioritizes accuracy, neutrality, and adherence to the Neutral Point of View (NPOV) and Verifiable Sourcing policies. AI flags anomalies; humans adjudicate.
Top removal categories: Unverified claims (41%), promotional content (28%), harassment/vandalism (19%), copyright infringement (12%). All enforcement actions are logged in our public audit trail.
AI & Algorithmic Transparency
Aevum uses machine learning to enhance search relevance, detect citation drift, and surface cross-disciplinary connections. We maintain strict boundaries: AI never generates published content, and all algorithmic decisions are explainable.
- Search ranking weights are publicly documented (neutrality score: 40%, recency: 25%, citation depth: 35%)
- Recommendation systems exclude political, medical, and legal topics from automated curation
- Biannual third-party audits by independent AI ethics firms (results published)
- All training datasets are licensed, anonymized, and scrubbed of PII
🔍 Open Algorithm Initiative: We've open-sourced our citation verification engine under Apache 2.0. Researchers can inspect how source reliability scores are calculated.
Data Privacy & User Rights
We collect minimal data: IP addresses (hashed), read history (optional), and contribution metadata. No behavioral advertising, no tracking pixels, no third-party data brokers.
| Request Type | Received | Fulfilled | Avg. Response Time |
|---|---|---|---|
| Data Export (GDPR/CCPA) | 12,405 | 12,388 (99.9%) | 6.2 days |
| Account Deletion | 8,120 | 8,115 (99.9%) | 3.1 days |
| Correction / Update | 4,330 | 4,290 (99.1%) | 5.8 days |
| Consent Withdrawal | 1,890 | 1,890 (100%) | 1.4 days |
We comply with GDPR, CCPA, LGPD, and PIPL. All data is stored in encrypted, region-sharded clusters. Retention policy: 12 months for logs, 24 months for contributions, then anonymized or deleted.
Editorial Independence & Verification
Content integrity is our highest priority. Every article passes through a three-tier verification pipeline:
- Source Validation: AI scans for DOI/ISBN/URL validity, domain reputation, and publication date
- Expert Review: Domain specialists verify claims, context, and neutrality
- Consensus Check: Cross-referenced against 14+ academic databases and peer-reviewed journals
In 2024, we rejected 3,210 editorial overrides requested by external parties. 0% of articles were altered under non-judicial pressure. Our Independent Editorial Council (12 members from 8 countries) publishes quarterly integrity assessments.
Methodology & Limitations
Data covers Jan 1 – Dec 31, 2024. Figures are compiled from internal moderation logs, legal request trackers, and automated privacy dashboards. Third-party audits were conducted by OpenMetric Labs and Veritas Academic Group. Some jurisdictions restrict public disclosure; we provide aggregated ranges where exact numbers cannot be shared. Definitions and calculation methods are available in our Transparency Methodology Appendix.
Next Steps & Contact
In 2025, we will: launch a real-time transparency dashboard, expand AI audit frequency to quarterly, publish contributor demographic breakdowns, and integrate a public challenge portal for disputed articles.
Questions, corrections, or submissions? Contact our transparency team at transparency@aevumenc.org. Report content concerns via Report Center. This report is licensed under CC BY-SA 4.0.