Skip to content
#explainable ai Open access

JAZB Framework v1.3: Judiciary AI Zero-Trust Broker

Oct 2026 · Zenodo (CERN European Organization for Nuclear Research)

Abstract

JAZB (Judiciary AI Zero-Trust Broker) is an Authority-centric, human-sovereign enterprise governance framework and architecture for artificial intelligence designed to govern the establishment, delegation, interpretation, exercise, assurance, and revocation of organizational Authority. JAZB applies Zero Trust, least privilege, separation of duties, jurisdictional separation, Human-Established Law, bounded autonomy, accountable decision rights, and continuous governance assurance to AI-enabled operations. The framework distinguishes technical capability from legitimate Authority and establishes an architecture in which humans remain the source of organizational Law while artificial intelligence may interpret and exercise bounded Authority without becoming the source of its own Authority. JAZB separates enterprise governance from machine capability. Business purpose, human decision rights, Law establishment, architecture, residual-risk acceptance, technical operation, expansion of autonomy, and independent assurance are treated as distinct governance responsibilities. Technical access, credentials, network reachability, model capability, workflow ownership, or administrative control do not independently establish legitimate Authority. JAZB defines separate Internal and Border Jurisdictions. The Judge serves as the AI interpretation and decision layer for Internal Jurisdiction, while Enterprise AI Customs governs lawful passage across the Organization’s external Organizational Boundary. These jurisdictions are supported by Human-Established Law Layers, Commissioners, operational AI AGENTs, Port Authority, Port Directors, Customs AGENTs, Judicial Instruments, independent Actuation Authority, the Chain of Authority, the Chain of Evidence, emergency governance, Governance Escape Resistance, trusted recovery, and explicit Law-establishment procedures. The framework is designed to complement established enterprise architecture, cybersecurity, identity, risk-management, privacy, Zero Trust, AI-governance, service-management, and management-system practices rather than replace existing technical controls or standards. Version 1.3 is a framework and architecture clarity release. It preserves the human-sovereign Authority model, dual-jurisdiction architecture, behavioral governance, bounded execution, enterprise operating model, and continuous assurance principles established through Version 1.2 while making the relationship among organizational purpose, enterprise architecture, governance functions, and technical implementation explicit. Version 1.3 formalizes four related but non-equivalent levels: JAZB Framework → Enterprise Authority Architecture → Authority Governance Functions → Technical Realization The JAZB Framework defines purpose, doctrine, stakeholders, desired outcomes, governance, lifecycle, and assurance. The Enterprise Authority Architecture relates those concerns into a coherent model for legitimate machine Authority. Authority Governance Functions preserve the required relationships during operation. Technical Realization uses enterprise technologies to implement those functions without becoming the source of organizational Law, Authority, or sovereignty. This hierarchy makes JAZB easier to examine from executive, architectural, governance, engineering, operational, and assurance perspectives. A product, gateway, model, policy engine, workflow, or control platform does not become the entire JAZB Framework merely because it implements one or more JAZB functions. Version 1.3 also formalizes an Architectural Traceability Spine: Organizational Purpose → Stakeholder Concerns and Constraints → Governance Decision → Human-Established Law → Bounded Authority → Governed Action → Evidence → Outcome → Reassessment Material governed Scope should remain traceable through this sequence. Architecture and assurance records should explain why an AI capability is being used, whose concerns and obligations affect that use, who holds the relevant decision rights, how legitimate Authority is established, how consequential action is governed, and what evidence supports continued operation. The release defines six complementary canonical architecture views: Enterprise Context; Governance and Decision Rights; Authority Architecture; Operational and Jurisdiction; Evidence and Assurance; and Lifecycle and Change. These views describe one governed architecture while exposing the concerns relevant to different stakeholders. They do not create competing sources of Law, Authority, jurisdiction, or evidence. Version 1.3 strengthens Current State, Target State, and transition integrity. Planned capabilities, proposed architecture, and intended controls must remain distinguishable from implemented and validated governance. Material architecture decisions and transitions should preserve accountable ownership, significant alternatives, dependencies, assumptions, known limitations, rollback or exit conditions, evidence gates, and reassessment triggers. The release also makes Technical Realization mapping explicit. Material products, services, platforms, and controls should be mapped to the Authority Governance Functions and architecture properties they realize or support. Where one platform performs multiple JAZB functions, required logical separation, decision rights, Authority boundaries, and evidence independence must remain effective. Version 1.3 expands outcome, portfolio, and economic governance. Architecture and adoption decisions connect intended business or service outcomes with governance outcomes, evidence, support burden, integration complexity, human-review demand, provider dependency, recovery requirements, reversibility, and expected cost at scale. Economic pressure, implementation convenience, and technical availability do not create or expand machine Authority. Continuous Governance Assurance remains foundational. Governance is not permanently established by policy, configuration, deployment, certification, or prior testing. JAZB retains the repeating assurance cycle: Establish → Operate → Observe → Validate → Discover → Remediate → Revalidate Material change may invalidate previously valid assumptions and require Authority, architecture, enforcement, or assurance reevaluation. A successful prior test does not constitute permanent assurance. Technical recovery does not automatically restore prior Authority. Missing evidence, stale assurance, unknown state, provider opacity, or an unmediated consequential path must not be represented as positive proof that governance is operating correctly. Material residual-risk acceptance remains a human governance responsibility. Prevention and containment claims remain bounded to paths actually mediated by the required governance or enforcement architecture, or by equivalent controls capable of reliably preventing the consequential action. Known gaps, uncertain attribution, missing telemetry, exceptions, third-party limitations, and unresolved dependencies must remain visible in the assurance record. Version 1.3 makes standards and compliance mapping architecture-aware. Formal crosswalks continue to distinguish Direct Alignment, Complementary Alignment, Implementation Support, and No Equivalence while identifying which JAZB architectural level a relationship supports. Technical implementation support does not independently establish framework-level governance equivalence, and architectural alignment does not prove that a control is deployed or effectively governing consequential action. This release contains the seven core JAZB Framework publications and an informational, non-normative release-notes companion: Publication 1 | Phase I: Foundational ArchitectureEstablishes the JAZB Framework, Enterprise Authority Architecture, foundational doctrine, human sovereignty model, organizational decision rights, Authority Governance Lifecycle, Human-Established Law, Internal and Border Jurisdictions, and continuous governance assurance. Version 1.3 makes the four-level hierarchy, stakeholder concerns, six canonical architecture views, architectural traceability, progressive disclosure, and outcome linkage explicit. Publication 2 | Phase II: Authority & Judicial SpecificationFormalizes Rights, Duties, Liberty, Authority, Delegation, Baseline and Elevated Authority, Point Policy, judicial decision semantics, Judicial Instruments, independent Actuation Authority, effective-target and method governance, the Chain of Authority, the Chain of Evidence, and internal Authority workflows. Version 1.3 positions these functions within the broader framework and architecture, showing how organizational purpose and human decision rights become bounded internal machine Authority. Publication 3 | Phase III: Enterprise AI Customs SpecificationDefines Border Jurisdiction, lawful ingress and egress, Port Authority, Port Director, Customs AGENT, Ports of Entry, governed passage, cross-boundary delegation, inspection, enforcement, external-service ownership, provider change, third-party opacity, and border Governance Escape Resistance. Version 1.3 adds stakeholder and external-relationship concerns, border architecture views, and traceability from organizational decisions through Border Law, passage, enforcement, evidence, and reassessment. Publication 4 | Phase IV: Security & Resilience SpecificationDefines compromise handling, Governance Escape Resistance, Law-state integrity, evidence integrity, governance continuity, degraded operation, dependency failure, trusted recovery, recovery reauthorization, emergency resilience, and continuous assurance under failure and change. Version 1.3 clarifies resilience as an enterprise architectural property and connects recovery and assurance decisions to organizational outcomes, stakeholder concerns, obligations, dependencies, and architecture state. Publication 5 | Phase V: Implementation & Adoption GuideProvides the enterprise operating model for

View source

Similar papers

#artificial intelligence Conference Open access Apr 2020

ECCOLA - a Method for Implementing Ethically Aligned AI Systems

The method, ECCOLA, is presented, which aims at making the high-level AI ethics principles more practical, making it possible for developers to more easily implement them in practice.

Ville Vakkuri, Kai-Kristian Kemell, P. Abrahamsson · 64 citations · ⚡6
#computer vision Review Apr 2024

AI-powered Code Review with LLMs: Early Results

The goal is to not only refine the accuracy of the LLM-based tool but also to underscore its potential in streamlining the software development lifecycle through proactive code improvement and education.

Z. Rasheed, Malik Abdul Sami, Muhammad Waseem et al. · 62 citations · ⚡3
#computer vision Open access Mar 2024

LLM-based agents for automating the enhancement of user story quality: An early report

The use of large language models to automatically improve the user story quality in Austrian Post Group IT agile teams is explored, with a reference model for an Autonomous LLM-based Agent System developed and implemented at the company.

Zheying Zhang, M. Rayhan, Tomas Herda et al. · 48 citations · ⚡4
#computer vision Review Mar 2024

System for systematic literature review using multiple AI agents: Concept and an empirical evaluation

This paper introduces a novel multi-AI-agent system designed to fully automate SLRs, and demonstrates how it substantially reduces the time and effort traditionally required for SLRs while maintaining comprehensiveness and precision.

Abdul Malik Sami, Z. Rasheed, Kai-Kristian Kemell et al. · 44 citations · ⚡2
#computer vision Feb 2024

Can Large Language Models Serve as Data Analysts? A Multi-Agent Assisted Approach for Qualitative Data Analysis

The proposed LLM-based multi-agent system automates qualitative data analysis process, creating opportunities for researchers and practitioners, and future improvements focus on enhancing multilingual performance and integrating continuous expert feedback.

Z. Rasheed, Muhammad Waseem, Aakash Ahmad et al. · 41 citations
#artificial intelligence Conference Open access Jun 2018

The Key Concepts of Ethics of Artificial Intelligence

It is suggested that the focus on finding keywords is the first step in guiding and providing direction for future research in the AI ethics field.

Ville Vakkuri, P. Abrahamsson · 39 citations · ⚡2

Related blog posts

Microsoft Research Blog Oct 7, 2026

Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses

Training AI agents with reinforcement learning can be challenging because their tools, context, and decision-making are managed by complex frameworks. Agent Lightning connects existing agents to RL training, making it easier to improve them without rebuilding them. The post Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.