Who Can Stop the Machine? The Executive Control Test for Deployed AI

Many organizations can approve the purchase of an artificial intelligence system but cannot govern it effectively after deployment. Legal responsibility may remain with the enterprise while technical control rests with a vendor. A designated human reviewer may be present but lack access to system logs, authority to suspend operations, or the practical ability to challenge an automated outcome. This talk introduces a six-part Executive Control Test for determining whether institutional authority over AI is real or merely nominal: Who can audit, challenge, pause, override, replace, and shut down the system? Through practical scenarios involving automated employment decisions, financial services, public administration, and agentic workflows, the session connects executive responsibility to procurement, contractual rights, system architecture, escalation procedures, incident response, and ongoing monitoring. Participants will leave with a practical framework for identifying gaps between formal accountability and actual control, assigning decision rights before deployment, and building intervention authority into an organization’s AI strategy and operating model.
Format: TRANSFORMATIONAL TALK
Location: ROOM C
Speakers: Arthur Sidney
Level: Intermediate
Audience: Registered participants