The Justice Automated Operational Governance Standard (JAOGS™) 2026-002 establishes the formal governance principles that define how execution-layer decision environments are documented, interpreted, and governed within AI-supported justice systems. As a companion to JAOGS™ 2026-001 (Foundation Doctrine), this document articulates the normative rules that ensure consistency, neutrality, and integrity in the application of Justice Decision Observability™ (JDO™). This standard addresses a critical and previously undefined governance gap: the space between automated system outputs and human authority, where operational decisions are made under conditions of uncertainty, workload pressure, and institutional constraint. While existing governance approaches focus on system performance, compliance, or policy design, JAOGS™ 2026-002 establishes a principled framework for documenting how decisions are actually made in practice. The standard defines a structured set of governance principles, including observability, reconstruction integrity, contextual grounding, non-adversarial documentation, consistency, measurement integrity, separation from system evaluation, authority visibility, and institutional neutrality. These principles are expressed using normative language to support consistent implementation across agencies, oversight bodies, and governance practitioners. JAOGS™ 2026-002 does not evaluate systems, assign fault, or determine compliance. Instead, it provides a neutral, structured foundation for documenting decision pathways and governance conditions at the point of execution. By formalizing these principles, the standard enables reliable reconstruction of decision events, strengthens institutional accountability, and supports defensible governance across high-stakes operational environments. This document is part of the broader Justice Decision Observability™ canonical framework and contributes to the establishment of a new governance discipline focused on the human decision layer in AI-supported justice systems.
Fleming et al. (Sun,) studied this question.