Irreducibly Human Series · Course 4 of 6 · Concept Brief · Graduate Engineering · Northeastern University

Conducting AI: the five supervisory capacities no algorithm possesses

What AI can and can't do — directing AI systems toward meaningful outcomes

Bear Brown & Company / Kindle Direct Publishing, 2026  ·  15-week semester
Concept Brief v0.1  ·  [Date]  ·  Reviewed by Dev the Dev
Document status — pre-TIC TOC

This is a concept brief, not a chapter specification document. It accurately represents the course at its current stage of development. The thesis, reader profile, three-act chapter map, and pedagogical architecture are documented here. Chapter specifications are pending.

This document is not ready for course delivery. The following must be completed before this reaches TIC TOC status:

Contents

  1. Thesis and Central Argument
  2. Reader Profile
  3. The Five Supervisory Capacities
  4. Three-Act Chapter Map
  5. Pedagogical Architecture
Section 1

Thesis and Central Argument

Thesis
AI solves faster than any human. Verification, reframing, interpretation, orchestration, and integration remain irreducibly human. These five capacities are now the decisive professional differentiator in AI-assisted knowledge work.

Training programs that build tool competence without building supervisory metacognition are producing performers when they need conductors.

Graduate engineers learn to use AI tools. They learn to prompt, delegate, and verify. Then they encounter the situation every employer now recognizes: something feels wrong and they cannot say why. The problem they have been handed is the wrong problem. The AI output is accurate, efficient, and pointed in the wrong direction.

The conductor metaphor unifies all five capacities: a conductor does not play any instrument — they hold the whole performance in their head, hear the wrong note before the score confirms it, and decide which piece is worth performing. The performance collapses without them even though they produce no sound themselves. As AI capability scales, the conductor role becomes more consequential, not less.

FieldValue
Series positionCourse 4 of 6 · Irreducibly Human: What AI Can and Can't Do
Deployment context15-week graduate course · College of Engineering · Northeastern University
Also appropriate forProfessional development; executive education
PrerequisiteAI tool competence — Course 1 of the series (AI Literacy, Fluency, and Trust) or equivalent professional experience
Section 2

Reader Profile

A second-year graduate engineer who prompts well, delegates confidently, and still freezes when an AI output feels wrong before she can prove it. She has never been asked to conduct. Only to play.

This reader can describe what went wrong after the fact. She cannot yet name it in the moment — which means she cannot stop it. The five capacities this course builds are the ones that turn the felt wrongness into a located, correctable judgment before the performance goes off the rails.

Section 3

The Five Supervisory Capacities

Each capacity is developed across two weeks: a framework week (peaks at Analyze) and an application week (peaks at Create or Evaluate). No student goes more than one week without an Apply-or-above deliverable.

Capacity 1 · Chapters 4–5
Plausibility Auditing

Hears the wrong note before the score confirms it. Evaluates AI output for coherence, domain fit, and hidden assumptions — before the output is acted on.

Capacity 2 · Chapters 6–7
Problem Formulation

Decides which problem is worth solving. Reframes the brief before generating. Identifies when the AI is solving the stated problem rather than the real one.

Capacity 3 · Chapters 8–9
Tool Orchestration

Knows when to bring in the brass. Selects, sequences, and combines AI tools based on what the problem actually requires — not habit or availability.

Capacity 4 · Chapters 10–11
Interpretive Judgment

Supplies the meaning the output cannot supply for itself. Translates AI findings into consequential decisions — knowing what the data shows and what it cannot show.

Capacity 5 · Chapters 12–13
Executive Integration

Holds the whole performance in their head. Synthesizes outputs across tools, timelines, and stakeholders into a coherent course of action the professional can defend.

Section 4

Three-Act Chapter Map

Status: Chapter titles and deliverable descriptions are concept-level only. Learning outcomes, opening strategies, worked examples, and bridge statements are pending. This map is a structural skeleton, not a specification.
Act One · Weeks 1–3 · Chapters 1–3

The Gap — diagnosing the conductor problem

Ch. 1 The Conductorless Orchestra Diagnoses the failure; names the gap
Ch. 2 The Solve-Verify Asymmetry Understands why the conductor role is structural, not temporary
Ch. 3 Three Levels of AI Usage Identifies their own supervisory gap
Act Two · Weeks 4–13 · Chapters 4–13

The Five Capacities — one per two-week block

Ch. 4–5 Plausibility Auditing Hears the wrong note before the score confirms it
Ch. 6–7 Problem Formulation Decides which problem is worth solving
Ch. 8–9 Tool Orchestration Knows when to bring in the brass
Ch. 10–11 Interpretive Judgment Supplies the meaning the output cannot supply for itself
Ch. 12–13 Executive Integration Holds the whole performance in their head
Act Three · Weeks 14–15 · Chapters 14–15

The Full Performance — adversarial assessment

Ch. 14 The Full Performance Peer critique and revision under pressure
Ch. 15 The Plausibility Audit Survives adversarial AI audit; names what the auditor cannot catch

Longitudinal cases

Three cases run across multiple chapters, producing a through-line that connects every capacity to a single arc of professional development. Case content is pending specification.

CaseChaptersStatus
Biomedical engineering analysis2, 4, 5, 11Named — content pending
Supervisory analysis13, 14, 15Named — content pending
Personal case inventory1, 15Named — content pending
Section 5

Pedagogical Architecture

Four structural features are identified as essential to the course. They are named and described at concept level. Prompt specifications, rubrics, and implementation protocols are pending.

Essential The Plausibility Audit

Claude functions as adversarial assessor in the final two weeks. The student's grade depends in part on what the auditor cannot find — proving development rather than self-report. The audit prompt architecture must be specified before the course runs: too much structural scaffolding and the audit tests prompt design, not supervisory capacity; too little and results are inconsistent across cohorts.

Pending: controlled prompt specification; rubric for what constitutes a passing audit; policy on what happens when the auditor finds nothing the student named.

Essential The Assessment Spine

Every deliverable requires the student to name, in writing, one judgment call that required their values, their domain knowledge, or their accountability that an AI could not have made on their behalf. Fifteen instances across the semester. Non-optional. A deliverable that cannot answer this question has not cleared the course's minimum threshold regardless of technical quality.

Pending: instance mapping across all 15 deliverables; rubric for what constitutes a specific vs. generic declaration.

Essential Longitudinal Cases

Three cases run across multiple chapters — the biomedical engineering analysis, the supervisory analysis, and the personal case inventory — producing a through-line that connects every capacity to a single arc of professional development. Cases are named but not yet written.

Pending: case content for all three longitudinal threads; worked examples per chapter; connection between case data and chapter deliverables.

Essential Framework/Application Structure

Every capacity is developed across two weeks — a framework week that peaks at Analyze and an application week that peaks at Create or Evaluate. No student goes more than one week without an Apply-or-above deliverable. This pacing constraint must hold across all ten capacity chapters.

Pending: Bloom's distribution map confirming the pacing constraint holds across all chapters; verification that Act One (three orientation chapters) does not break the one-week Apply-or-above rule.

Irreducibly Human: Conducting AI — Concept Brief v0.1 · [Date]
Pre-TIC TOC. Thesis, reader profile, chapter map, and pedagogical architecture are documented. Chapter specifications, learning outcomes, worked examples, and assessment rubrics are pending before this document reaches TIC TOC status.