Governance & Standards
Editorial Principles
The principles that govern what KARADAVI publishes, how content is authored and verified, and the strict boundaries on automation.
Knowledge is automatic.
The knowledge graph grows and synthesizes through automated extraction, entity resolution, and relation discovery.
Writing is intentional.
Editorial synthesis is authored, structured, and deliberately shaped by human editors and domain stewards.
Publishing is human.
No content enters the public knowledge graph without explicit, manual human review and sign-off in the CMS.
What KARADAVI Publishes
KARADAVI publishes structured knowledge about canonical entities: people, companies, technologies, scientific concepts, historical events, places, and ideas. Content is organised into two primary forms:
- Entity Profiles: Canonical, authoritative records that aggregate structured facts, relationships, and sourced claims about a specific subject.
- Editorial Articles: In-depth documents that explain, contextualise, or trace connections between entities and concepts across time.
The Role of AI
AI is infrastructure. It assists KARADAVI editorial operations in specific ways, and is explicitly prohibited from others. It never holds editorial authority.
AI May Assist With:
- ✓ Research and primary source discovery
- ✓ Entity and relationship extraction
- ✓ Category classification and taxonomy
- ✓ Draft synthesis for human review
AI May Not:
- ✕ Publish content autonomously
- ✕ Approve or authorize publication
- ✕ Alter live entity records autonomously
- ✕ Establish factual truth without sources
Verification & Corrections
Every claim in KARADAVI is bound to a verified primary source. If a published entity profile or article contains a factual error, it is reviewed and updated immediately upon verification of correcting evidence.
For detailed citation standards, see Source Policy. For a technical overview of our extraction engine, see How KARADAVI Works.