Core Problems FactHarbor Solves

Last modified by Robert Schaub on 2025/12/24 20:33

Core Problems FactHarbor Solves

Our Mission

FactHarbor brings clarity and transparency to a world full of unclear, contested, and misleading information by shedding light on the context, assumptions, and evidence behind claims.


Core Problems

Problem 1 — Misinformation & Manipulation

Falsehoods and distortions spread rapidly through:

  • Political propaganda
  • Social media amplification
  • Coordinated influence networks
  • AI-generated fake content

Users need a structured system that resists manipulation and makes reasoning transparent.

Problem 2 — Missing Context Behind Claims

Most claims change meaning drastically depending on:

  • Definitions
  • Assumptions
  • Boundaries
  • Interpretation

FactHarbor reveals and compares these variations.

Problem 3 — "Binary Fact Checks" Fail

Most fact-checking simplifies complex claims into:

  • True
  • Mostly True
  • False

This hides legitimate contextual differences.

FactHarbor replaces binary judgment with scenario-based, likelihood-driven evaluation.

Problem 4 — Good Evidence Is Hard to Find

High-quality evidence exists — but users often cannot:

  • Locate it
  • Assess its reliability
  • Understand how it fits into a scenario
  • Compare it with competing evidence

FactHarbor aggregates, assesses, and organizes evidence with full transparency.

Problem 5 — Claims Evolve Over Time

Research and understanding change:

  • New studies emerge
  • Old studies are retracted
  • Consensus shifts

FactHarbor provides:

  • Full entity versioning
  • Verdict timelines
  • Automatic re-evaluation when inputs change

Problem 6 — Users Cannot See Why People Disagree

People often assume others are ignorant or dishonest, when disagreements typically arise from:

  • Different definitions
  • Different implicit assumptions
  • Different evidence
  • Different contexts

FactHarbor exposes these underlying structures so disagreements become understandable, not divisive.


Core Concepts

Claim

A user- or AI-submitted statement whose meaning is often ambiguous and requires structured interpretation.

A claim does not receive a single verdict — it branches into scenarios that clarify its meaning.

Scenario

A structured interpretation that clarifies what the claim means under a specific set of:

  • Boundaries
  • Definitions
  • Assumptions
  • Contextual conditions

Multiple scenarios allow claims to be understood fairly and without political or ideological bias.

Evidence

Information that supports or contradicts a scenario.

Evidence includes empirical studies, experimental data, expert consensus, historical records, contextual background, and absence-of-evidence signals.

Evidence evolves through versioning and includes reliability assessment.

Verdict

A likelihood estimate for a claim within a specific scenario based on evidence quality, quantity, methodology, uncertainty factors, and comparison with competing scenarios.

Each verdict is versioned and includes a historical timeline.

AI Knowledge Extraction Layer (AKEL)

The AI subsystem that interprets claims, proposes scenario drafts, retrieves evidence, classifies sources, drafts verdicts, detects contradictions, and triggers re-evaluation when inputs change.

AKEL outputs follow risk-based publication model with quality gates and audit oversight.

Decentralized Federation Model

FactHarbor supports a decentralized, multi-node architecture where each node stores its own data and synchronizes via federation protocol.

This increases resilience, autonomy, and scalability.


Vision for Impact

FactHarbor aims to:

  • Reduce polarization by revealing legitimate grounds for disagreement
  • Combat misinformation by providing structured, transparent evaluation
  • Empower users to make informed judgments based on evidence
  • Support deliberative democracy by clarifying complex policy questions
  • Enable federated knowledge so no single entity controls the truth
  • Resist manipulation through transparent reasoning and quality oversight
  • Evolve with research by maintaining versioned, updatable knowledge

Related Pages