Specification

This section defines the technical architecture, data models, and functional requirements of FactHarbor.

1. Mission

FactHarbor brings clarity and transparency to a world full of unclear, controversial, and misleading information by shedding light on the context, assumptions, and evidence behind claims — empowering people to better understand and judge wisely.

2. Purpose

Modern society faces a deep informational crisis:

• Misinformation spreads faster than corrections.
• High-quality evidence is buried under noise.
• Meanings shift depending on context — but this is rarely made explicit.
• Users lack tools to understand *why* information conflicts.
• Claims are evaluated without clearly defined assumptions.
  FactHarbor introduces structure, transparency, and comparative reasoning. It provides:
• Multiple valid scenarios for ambiguous claims.
• Transparent assumptions, definitions, and boundaries.
• Full evidence provenance.
• Likelihood-based verdicts (one per scenario).
• Versioning and temporal change tracking.
• Hybrid AI–human collaboration.

3. Core Concepts

• Claim: A statement needing structured interpretation.
• Scenario: Definitions, assumptions, boundaries, and context.
• Evidence: Information supporting or contradicting a scenario.
• Verdict: Likelihood estimate based on weighted evidence for a specific scenario.
• Summary View: User-facing overview.
• AKEL: AI subsystem for drafting and assistance (human supervised).
• Federation: Decentralized nodes hosting datasets.
• Truth Landscape: The aggregation of multiple scenario-dependent verdicts showing where a claim is plausible.
• Time Evolution: Versioning of all entities allowing historical views.

4. Functional Lifecycle

The system follows a six-step lifecycle:

1. Claim submission: Automatic extraction and normalisation; Cluster detection.
   2. Scenario building: Clarifying definitions and assumptions; AI proposals with human approval.
   3. Evidence handling: AI-assisted retrieval; Human assessment of reliability; Explicit scenario linking.
   4. Verdict creation: AI-generated draft verdicts; Human refinement; Reasoning explanations.
   5. Public presentation: Concise summaries; Truth Landscape comparison; Deep dives.
   6. Time evolution: Versioning of all entities; Re-evaluation triggers when evidence changes.

5. Chapters

This specification is organized into the following sections:

Core Documentation

• AI Knowledge Extraction Layer (AKEL) - AI system architecture
• Architecture - System architecture and design
• Automation - Automated processes and workflows
• Data Model - Database schema and entities
• Requirements - Functional and non-functional requirements
• Workflows - Process workflows

Diagrams

• Diagrams - Visual architecture and workflow diagrams

Additional Resources

• Data Examples - Sample data structures
• FAQ - Frequently asked questions
• Federation & Decentralization - Future federation plans
• Review & Data Use - Data usage policies