At 2:10 AM over the Atlantic, an autopilot handed control back to two pilots with a soft chime. Four minutes later, Air France 447 was in the ocean. The investigation didn't blame the autopilot. It blamed the handover. We are about to deploy thousands of AI agents into hospitals, courts, tax authorities, and citizen portals. One day — maybe not today, maybe not next year — the chime will sound in your organization too. Three questions will arrive together: Who is running it? Who can stop it? Who is accountable for it? In my new article, I share The Oversight Matrix — a simple framework I've been using with public sector leaders to decide where the human should stand for every AI decision: in the loop, on the loop, or out of it. If you can't name the human accountable for one AI system in your organization today — this is for you. Read here
