Core Editorial Principles
All contributions must align with these foundational standards that ensure accuracy, reliability, and editorial integrity across the platform.
Encyclopedic Accuracy
Claims must be factually correct, up-to-date, and free from speculation or unverified assertions.
Neutral Point of View
Present multiple perspectives fairly. Avoid loaded language, advocacy, or subjective framing.
Verifiability
Every non-trivial claim requires attribution to a reliable, independently published source.
Notability
Topics must demonstrate significant coverage by reliable secondary sources to warrant standalone coverage.
Structure & Formatting
Article Anatomy
Standard entries follow a consistent structure to ensure readability and information architecture:
- Lead Section: 2–4 sentences defining the topic, its significance, and key facts. Must be self-contained.
- Body Sections: Logical progression (History → Development → Applications → Impact → Criticism → Legacy)
- Infobox: Structured data sidebar for quick reference (auto-generated from structured fields)
- References: Formatted citations using the platform's citation manager
- External Links & Further Reading: Curated, non-promotional resources only
Media & Visuals
Images, diagrams, and maps must be properly licensed (CC BY-SA 4.0 or public domain), include alt text, and be credited in the description field. Vector graphics are preferred for technical illustrations.
Citation Standards
References are the backbone of Aevum's credibility. Follow these requirements:
- Use peer-reviewed journals, academic books, major news outlets, and official institutional publications.
- Self-published sources, forums, social media, and Wikipedia are not acceptable as primary citations.
- Include author(s), title, publisher, date, and stable URL/DOI.
- For digital sources, archive URLs (e.g., Wayback Machine) are required if the original may become inaccessible.
Smith, J. & Lee, A. (2023). "Advances in Neural Architecture Search." Journal of Machine Learning Research, 24(12), 301–328. doi:10.5555/jmlr.2023.301
AI & Automation Policy
Aevum leverages AI for discovery and structuring, but human editorial oversight remains mandatory.
- AI-Assisted Drafting: Permitted for grammar, structuring, and initial research synthesis. Must be disclosed in edit summary as "AI-assisted".
- Full AI Generation: Prohibited for standalone submissions. AI outputs require line-by-line human verification and source mapping.
- Algorithmic Edits: Bots and automation scripts must be registered, sandboxed, and approved by the moderation team before deployment.
Our AI systems flag low-confidence claims, potential bias, and missing citations for human review. Final editorial authority always rests with verified contributors and subject editors.
Notability & Scope
Not every topic deserves a dedicated article. Use this framework to evaluate scope:
- Individuals: Must have significant coverage in multiple independent, reliable sources. Self-published biographies or press releases do not qualify.
- Organizations: Must demonstrate impact, longevity, or recognition in their field. Local clubs or startups without external coverage should be merged or omitted.
- Events: Must have historical, scientific, or cultural significance documented by secondary sources.
- Concepts/Theories: Must be established in academic or professional discourse with clear definitions and applications.
When in doubt, propose the topic in the community forum for consensus before drafting.
Review & Moderation Process
Every edit passes through a transparent, multi-stage pipeline:
- Submission: Edit is queued with change summary and source links.
- AI Pre-Check: Automated scan for formatting, citation validity, and policy violations.
- Peer Review: Assigned to subject-matter contributors or trained editors.
- Publish/Revise: Approved edits go live. Flagged changes return to author with detailed feedback.
Disputes can be appealed via the editorial dashboard. Repeated violations of guidelines may result in contributor probation or account review.