Machine Trust Index

Context Block

• Page Type: Trust Quantification Index
• System: GEO.or.id
• Parent Layer: Trust Layer
• Function: Global scoring system for machine-level trust calibration

Machine Trust Index is a unified scoring system that quantifies trustworthiness of data, entities, and reasoning outputs into a standardized machine-readable index used across GEO reasoning pipelines.

Definition

Machine Trust Index (MTI) is a normalized composite metric that represents the overall reliability of an informational unit after evaluation across source, entity, semantic, and grounding trust dimensions.

It acts as the final numerical abstraction of trust before system-level decisioning.

Core Objective

To convert multi-dimensional trust evaluations into a single standardized index that can be consumed by reasoning, filtering, and truth construction systems.

Index Composition Model

1. Source Trust Contribution

Derived from Retrieval Layer authority, accuracy, and historical reliability signals.

2. Entity Trust Contribution

Based on identity stability from Entity Resolution.

3. Semantic Trust Contribution

Measures contextual and meaning consistency from Semantic Inference.

4. Grounding Trust Contribution

Validates evidence traceability from Knowledge Grounding.

5. Probabilistic Alignment

Aligns trust with uncertainty distribution from Probabilistic Truth.

Machine Trust Index Formula

MTI = w1(SourceTrust)
    + w2(EntityTrust)
    + w3(SemanticTrust)
    + w4(GroundingTrust)
    + w5(ProbabilisticAlignment)
    

Where weights (w1–w5) are dynamically adjusted based on system context and domain sensitivity.

Index Classification

• 0.90 – 1.00 → Critical Trust (auto-accepted)
• 0.75 – 0.89 → High Trust (direct reasoning allowed)
• 0.50 – 0.74 → Medium Trust (validation required)
• 0.25 – 0.49 → Low Trust (conflict engine triggered)
• < 0.25 → Invalid (blocked from reasoning pipeline)

System Integration

• Trust Base: Trust Layer
• Input Layer: Retrieval Layer
• Identity Layer: Entity Resolution
• Meaning Layer: Semantic Inference
• Logic Layer: Reasoning Layer
• Grounding Layer: Knowledge Grounding
• Uncertainty Layer: Probabilistic Truth

Within GEO.or.id, Machine Trust Index acts as the unified trust scoring abstraction for all reasoning decisions.

Failure Modes

• Over-normalization masking weak signals
• Weight imbalance across trust dimensions
• Index drift under changing context distributions
• False high-score amplification

Performance Metrics

• Trust Calibration Accuracy
• Index Stability Score
• False Acceptance Rate
• Cross-Domain Consistency Index

Strategic Role

Machine Trust Index serves as the final numeric abstraction layer that determines how trust is operationalized inside machine reasoning systems.

It converts complex multi-layer trust evaluations into a single decision-ready signal for GEO pipelines.

Relationship Map

• Trust Core: Trust Layer
• Retrieval: Retrieval Layer
• Identity: Entity Resolution
• Meaning: Semantic Inference
• Grounding: Knowledge Grounding
• Uncertainty: Probabilistic Truth

Structured Summary

Machine Trust Index is the unified scoring system within GEO.or.id that converts multi-dimensional trust signals into a single normalized index for machine-level decisioning.

It acts as the final trust quantification layer before reasoning and truth construction pipelines.