AI Risk – a User’s Guide (Part I): The Approval Nobody Signed off
Summary
Sanjeev Sharma argues that the riskiest AI decision this year may be one that nobody approved, highlighting a governance gap as AI systems operate autonomously. He emphasizes that governance must be continuous and parallel to deployment, not a single preflight check, and distinguishes governance failure from technology failure. The piece introduces the concept of Shadow AI and outlines a starting point for leaders: inventory every AI system, assign an owner and risk tier, and establish monitoring before harm occurs.