Markov Blanket Boundaries

Subsystem isolation and information flow visualization

HUMAN Human Subsystem

GOVERNANCE Governance Subsystem

AI AI Subsystem

████ Protected Data

Three-Subsystem Architecture

Live Boundary Crossings Monitoring

Loading crossings...

DTO Inspector

What is a DTO? Data Transfer Objects (DTOs) are immutable data structures that cross Markov blanket boundaries. Each DTO defines exactly which fields are exposed to the receiving subsystem and which remain protected. This ensures subsystem isolation while enabling controlled communication.

Select DTO Type

Select a DTO to see what fields are included vs. protected.

Why Data is Protected

EFE Scores

Expected Free Energy scores (G values) are the AI's internal decision metric. Exposing them could enable gaming or manipulation of the allocation system. They are logged to Governance audit for compliance but never returned to API.

efe_ai: ████████

efe_hybrid: ████████

efe_human: ████████

Model Parameters

Internal weights (beta values) and thresholds used for allocation decisions. Protected to maintain system integrity and prevent reverse-engineering.

beta_u: ████

beta_n: ████

thresholds: ████████

What You DO See

Categorical levels (low/medium/high/critical) and explanations are always provided. These enable informed human oversight without exposing internal mechanics.

uncertainty: "medium"

novelty: "low"

explanation: "Based on..."

Choice Expansion Metrics Loading...

Von Foerster's ethical imperative: "Act always so as to increase the number of choices." These metrics track system health and detect dead-end patterns.

Effective Choice Score

Mode Balance

AI 33% Hybrid 34% Human 33%

Entropy: -- / 1.0

Dead-End Risk: --

Option Expansion

Higher = more decisions preserve multiple viable options

Override Exercise

Rate of human override usage