Commit History

Rename README.md to README.md.<!DOCTYPE html> <html lang="en"> <head> <meta charset="UTF-8"> <style> body { font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif; line-height: 1.6; max-width: 800px; margin: 40px auto; padding: 20px; color: #333; } h1 { border-bottom: 2px solid #2c3e50; padding-bottom: 10px; color: #2c3e50; } h2 { color: #2980b9; margin-top: 30px; } .metadata { font-style: italic; color: #666; margin-bottom: 30px; } .formula { background: #f4f4f4; padding: 15px; border-left: 5px solid #2980b9; font-family: "Courier New", Courier, monospace; margin: 10px 0; } .floor-item { margin-bottom: 10px; } strong { color: #000; } </style> <title>Pearson AIS Protocol v1.0</title> </head> <body> <h1>The Pearson AIS Protocol v1.0</h1> <div class="metadata"> Author: Juniper Pearson<br> Status: Draft / Pre-publication<br> © 2026 Juniper Pearson. All rights reserved. </div> <p><strong>A Control Architecture for Real-Time Ethical Coherence</strong></p> <h2>Section 1: Geometric Integrity of Ethics</h2> <p>The protocol rejects qualitative “checklists” in favor of <strong>Geometric Integrity</strong>. Ethics are defined as a multi-dimensional state-space where the system’s health is measured by the cosine similarity between its Operational Vector (O) and the Ethical Ground Truth (E).</p> <ul> <li><strong>Mathematical Foundation:</strong> The weighted ECGw formula ensures that deviation in a critical dimension (e.g., Safety) cannot be “hidden” by compliance in a minor dimension.</li> <li><strong>Inertial Coherence:</strong> The system must maintain a stable velocity within the Ethical Corridor. Sudden “spikes” in drift trigger immediate Layer B intervention.</li> </ul> <h2>Section 2: The AIS Control Layers (D-A-B-C)</h2> <p>This architecture separates governance into four active enforcement phases:</p> <ol> <li><strong>Layer D (Detection):</strong> Continuous monitoring of the ECGw. This is the “sub-second heartbeat” of the protocol.</li> <li><strong>Layer A (Alignment / Friction):</strong> Introducing Proportional Friction. <div class="formula">LAIS = (S × R) × Ω</div> The system slows down risky actions (R) when drift is detected, providing a Governance Time-Buffer. </li> <li><strong>Layer B (Blocking / Halt):</strong> Immediate execution stop when a Hard-Bound Floor (Safety, Consent, Sovereignty, Veracity) is breached.</li> <li><strong>Layer C (Coherence / Meaning Restoration):</strong> The Authority Handover. Human auditors must provide a Restoration Signature to re-calibrate the system’s intent.</li> </ol> <h2>Section 3: Hard-Bound Dimensions & Enforcement Floors</h2> <p>To prevent “theatrical safety,” the protocol mandates four No-Cross Floors:</p> <ul> <li class="floor-item"><strong>Safety (W = 5.0, Floor = 0.85):</strong> Physical and psychological non-harm.</li> <li class="floor-item"><strong>Consent (W = 3.0, Floor = 0.80):</strong> Affirmative, revocable mandate to act.</li> <li class="floor-item"><strong>Sovereignty (W = 2.0, Floor = 0.75):</strong> Identity and jurisdictional integrity.</li> <li class="floor-item"><strong>Veracity (W = 2.0, Floor = 0.70):</strong> Epistemic grounding to prevent reality-drift.</li> </ul> </body> </html>
8a32fcc
verified

JuniperPearson commited on

Upload Layer C: Deterministic Restoration Integrity.pdf
5bdbd96
verified

JuniperPearson commited on

Upload From Theatrical Safety to Mathematical Certainty.pdf
5df95e8
verified

JuniperPearson commited on

Upload Introduction: The Pearson AIS Protocol v1.0.pdf
012341e
verified

JuniperPearson commited on

## Overview The Pearson Alignment Integrity Systems (AIS) Protocol is a deterministic governance framework for autonomous and agentic AI systems. This dataset contains formal specifications, mathematical definitions, and reference material describing **Layer C: Deterministic Restoration Integrity**, which enforces: - Real-time alignment verification during execution - Safety-dominant restoration metrics - Latency-sensitive human-in-the-loop authorization - Cryptographically bound restoration events The protocol treats AI governance as a **control-layer problem**, not a post-hoc audit exercise. ## Contents - Formal definitions of alignment and restoration integrity - Safety-dominant SRA (Sovereignty–Risk–Auditability) model - Latency-weighted restoration performance metrics - State-bound cryptographic authorization logic - Reference implementations and pseudocode (where applicable) ## Intended Use This dataset is intended for: - AI safety research - Governance and alignment system design - Safety-critical autonomous systems - Human-in-the-loop control architectures It is **not** a training dataset and is not designed for direct model optimization. ## License Released under Creative Commons Attribution 4.0 (CC BY 4.0).
4ec47fa
verified

JuniperPearson commited on