The Judgment Economy Has an Org Chart Now: Deciding Which Calls AI Makes and Which People Own

AI should take the triage. People should keep the judgment. The harder question for leaders is who owns the decisions AI now shapes.

There is a popular argument that AI will sort people by their willingness to do hard thinking. Those who use AI as a sparring partner pull ahead. Those who use it to avoid thinking fall behind.

It is true. It is also too abstract to act on. Security and operations leaders need something more specific: a map of where judgment is required and who holds it.

What AI should take

Incident management is a useful test case because the work is well documented. AI performs well at watching many signals at once, correlating alerts, executing mechanical first-response steps, and drafting the timeline for a postmortem. Most of that work is triage. Handing it to AI can relieve the alert fatigue that wears down human teams.

What people should keep

Declaring severity. Communicating with customers and the public. Making calls with a large blast radius. Owning the postmortem. AI can draft the status update. A person approves it and answers for it.

The real risk is skill decay

AI will occasionally get easy calls wrong, and that is manageable. The larger danger is that relying on AI for the easy calls erodes the skills needed for the hard ones. Junior staff who never troubleshoot manually never build the instinct. Teams that depend on AI tooling struggle when it fails. And AI is strongest on known patterns and weakest on novel failures, which is exactly where it can sound most confident while being wrong.

A three-question filter

  1. Is there a consistent right answer? If the correct response depends on context that changes, keep a person on it.
  2. What is the blast radius if the AI is wrong? Large blast radius means human ownership.
  3. Does a specific person have to answer for the outcome? If yes, that person owns the decision.

A task that passes all three can be automated. A task that fails any one stays with a human.

Where this sits in the Four Signals Framework

This is an Ownership question. The Ownership signal asks who is accountable for each AI system in production, and whether they know it. Most organizations can name the owner of the system. Far fewer can name the owner of each decision the system now influences. That second list is the one that matters when something goes wrong.

Questions for the board

Operational Takeaway

Ask your team to name the last five AI recommendations they acted on, who approved each one, and why. If that is hard to answer, you have a judgment ownership problem, and better tooling will not fix it.

An earlier version of this piece appeared in Northern Signal, Tim Reed's briefing on AI governance and security.


Tim Reed, CPP

Tim Reed, CPP, is the founder of The Reed Group and author of Signals in the Noise: Security, Technology, and the Hidden Patterns of Modern Risk (ISZ Press, 2026). With 28 years of experience across corporate security, executive protection, and intelligence programs at Apple and Uber, he built the firm to bring operational security judgment to AI governance and board advisory work. He writes at Northern Signal.

Who Owns the Decisions AI Shapes?

The AI Readiness Review identifies who is accountable for each AI system and the decisions it informs, and rates your Ownership posture as Governed, Exposed, or Ungoverned.

Request an AI Readiness Review