Test
Test
Blog / Articles / Perspectives
Think clearly. Question deeply.
Essays on AI governance, operational risk, and the systems that shape decisions.
19 essays
Decommissioning AI: Why Switching a System Off Is a Governance Decision
Switching off an AI system can take one click. Retiring it well takes a plan. What NIST's AI RMF (GOVERN 1.7, MANAGE 2.4) and the EU AI Act…
Rubber-Stamp Oversight: When a Named Reviewer Stops Deciding
An AI approval can name a human reviewer and still leave the system making the decision. What the EU AI Act (Articles 14 and 26) and NIST AI…
Silent Model Changes: When an AI Approval Outlives the Model It Covered
An AI approval covers a specific configuration: a model, its instructions and its data, tested at one point in time. When the model changes and the product name…
Intended-Use Drift: The AI Governance Risk Hidden Inside an Approved System
An AI approval is a judgment about one use: its users, data, decisions and potential harm. When an approved system quietly takes on a new job, that judgment…
The AI Inventory: The Governance Control Most Organizations Underrate
Every AI governance control, from risk assessment to monitoring, is applied to a specific system. Systems missing from the list never reach those controls. NIST's AI RMF, ISO's…
Governance Is a Cycle, Not a Gate: What ISO/IEC 42001 Asks of AI Oversight
An AI system's approval records what was true on the day it was signed. ISO/IEC 42001, the international standard for AI management systems, asks for more: a management…
AI Governance Foundations: Principles, Accountability, and Proportionate Control
A practical introduction to AI governance principles, including accountability, transparency, safety, fairness, privacy, and human oversight.
Public AI Governance: Regulation, Institutions, and International Coordination
How public institutions can govern AI through risk-based regulation, standards, enforcement, public participation, and international coordination.
An Operating Playbook for Responsible AI: From Inventory to Oversight
A practical operating model for organizations: inventory AI systems, classify risk, assign ownership, embed lifecycle controls, monitor performance, and prepare for audits.