⚡ Quick Summary
The AI Governance Practitioner’s Manual by Noah M. Kenney offers an operational, role-oriented blueprint for implementing artificial intelligence governance across enterprise ecosystems. Rejecting traditional straight-line governance—annual audits, generic principles, and advisory committees—the manual introduces the five-layer AI Governance Stack (Data Governance, Model Governance, System Integration, Control & Monitoring, and Audit & Evidence).
Through rigorous implementation specifications (DG-1 through AE-5), the manual maps emerging horizontal regulations (such as the EU AI Act), sector-specific rules (FDA SaMD, SR 11-7, CFPB Circulars), and state privacy/AI regimes into actionable engineering controls, assigned ownership matrices, and audit-ready evidence streams.
🧩 What's Covered
Structured across 46 chapters, nine parts, and ten reference appendices, the manual provides a modular reference architecture covering:
- The AI Governance Stack: Complete specifications for five interdependent operational layers: Layer 1 (Data: inventory, bias, quality, provenance), Layer 2 (Model: architecture, fairness testing, robustness, interpretability), Layer 3 (System Integration: pipeline security, circuit breakers, human-AI interaction), Layer 4 (Control & Monitoring: access control, real-time drift, anomaly detection), and Layer 5 (Audit & Evidence: immutable logs, documentation standards, regulatory reporting).
- The Cascading Failure Principle: A foundational rule demonstrating that governance failures cascade upward from data to model to integration, requiring remediation at the root layer rather than via superficial operational patches.
- Single-Owner Governance Architecture: Concrete accountability models replacing diffuse committee oversight with designated primary owners possessing unilateral authority to halt training, block releases, or trigger rollbacks.
- Global Regulatory Regimes: In-depth statutory analyses and Stack mappings for the EU AI Act, GDPR, EU Digital Regulation Suite (NIS2, DORA, CRA, Data Act), US Federal frameworks (FTC Act Section 5, FCRA, ECOA, HIPAA, CISA/CIRCIA, SEC), US State statutes (California AB 2013/SB 942/SB 53, Colorado CAIA, Texas TRAIGA, NYC Local Law 144, Illinois BIPA/HB 3773), and international frameworks across the UK, Americas, and APAC (China’s PIPL and AI Measures, Korea’s AI Basic Act, Singapore’s AI Verify).
- Cross-Sector Playbooks & Economics: Domain-specific governance implementations for healthcare, financial services, critical infrastructure, and autonomous systems, paired with a 12-month phased implementation roadmap and a comprehensive business case quantifying the cost of inaction and proactive governance ROI.
💡 Why it matters?
Most AI governance initiatives fail because they treat governance as an annual checklist or an abstract ethics statement. Kenney bridges the divide between legal interpretation and technical implementation. By translating statutory triggers into system-level RFC 2119 requirements and assigning discrete accountability per Stack layer, this resource turns compliance from an unpredictable legal liability into a repeatable, scalable engineering discipline.
❓ What's Missing
The excerpted chapters prioritize structural, legal, and operational controls over code-level script implementations or hands-on tutorials for specific statistical fairness toolkits. In addition, coverage reflects legislation and regulatory enforcement through August 2026, meaning emerging judicial challenges (such as post-2026 rulings on transatlantic data transfers or state AI preemption battles) will require continuous monitoring.
👥 Best For
Chief AI Officers, General Counsel, Data Protection Officers, Chief Information Security Officers, compliance managers, and ML engineering leads who need an actionable, audit-ready operational framework for governing enterprise AI across global jurisdictions.
📄 Source Details
- Author: Noah M. Kenney (AIGP, CIPM, M.Eng.)
- Publisher / Organization: Digital 520
- Edition / Publication Year: First Edition, 2026 (Coverage current through August 2026)
- Format: Reference Manual / Book (46 Chapters, Appendices A–J)
📝 Thanks to
Reviewed by Kuba Szarmach for the AI Governance Library (aigl.blog).
