AIEM theory walkthroughStep 1 of 13: Why a model
Need a different explanation?
Ask the AIEM tutor, or clear this short walkthrough and begin again.
AI Engagement Model · interactive theory walkthrough
Step 1 of 13
Keep your judgment in the loop
AI can hand you an answer or help you do better thinking. The difference is where your judgment enters the work.
About 20–30 minutes · 13 named steps · research detail is optional
The same person. The same task. Two workflows.
Where does human judgment enter the work?
Select any card to see what happens at that point in the workflow.
Answer handoff
AI as a vending machine
One request produces one polished output. The workflow may be efficient, but the person's criteria, checking, and final reasoning stay mostly invisible.
This can fit a bounded, low-stakes task. It becomes risky when the task requires learning, originality, evidence, or a consequential decision.
Judgment loop
AI as a workshop
AI still contributes substantial work. The person defines the job, examines alternatives, tests weaknesses, checks evidence, and decides what leaves the bench.
The sequence can change with the task. What matters is that judgment appears where the stakes require it.
Judgment loop
Frame: Define the real job
The person names the goal, constraints, audience, and decision criteria before asking AI to solve anything. This keeps the task from being defined by the first answer.
Imagine two people facing the same task. One asks AI for an answer, accepts it, and submits it. The other defines the goal, asks for options, tests the weak points, checks the evidence, and makes the final choice.
The difference is not whether AI produced something. It is where human judgment entered the work. AIEM gives you a vocabulary for seeing that difference and changing the workflow while the work is still underway.
One living system, five parts
- 1
Mechanism
The person, AI tool, task, and context shape the interaction together.
- 2
Dispositions
The tendencies a person brings into the interaction.
- 3
Principles
The rules that guide what the person keeps, delegates, and checks.
- 4
Behaviors
The eight observable functions used during the work.
- 5
Outcomes
What repeated engagement may strengthen or weaken over time.
Becoming is the return cycle: repeated practice and reflection can feed later dispositions and choices. It is not a sixth part.
Five parts, one loop
Mechanism explains why the person, tool, task, and context work together. Dispositions are tendencies the person brings. Principles are the rules they try to follow. Behaviors are the observable functions their AI moves perform. Outcomes are what the engagement may strengthen or weaken.
Becoming is not a sixth part. It is the return loop: repeated engagements may change what the person brings to the next one.
Quick check
Which difference matters most when comparing the two workflows above?
Where does your judgment enter?
Think of one recent AI task. Did you frame the goal, compare options, challenge assumptions, verify evidence, and make the decision, or did the workflow end when the first usable answer appeared?
Which mental model are you using?
Most people show up with one of five ideas about what using AI even is. Pick the one closest to how you actually treat it.
The walkthrough now builds the model one part at a time, then shows what the current research supports and where the evidence is still incomplete. You can open the research detail in the evidence step when you want it instead of reading it before seeing the model.