The 25-minute run of show
Times are elapsed from the start. Choose two claims and protect the
last eight minutes for implications and the close.
0–3 minutes · Open and choose
Read together, then choose a claim.
First minute: Show the
opening reading. Read it aloud or
give the room about 45 seconds to read.
A student submits an excellent paper with help from an AI agent.
The agent helped define the problem and develop the argument. To
understand what the student learned, we need to hear which choices
they made and see how they respond when an assumption changes.
AI can help people do more demanding work. It can also take over
the decisions through which they learn. As educators, we have to
decide where learners need to exercise judgment and what would
show it has developed.
Adapted from
The Irreducible Officer.
Fictional education example.
Second minute — ask: “Where have you seen the
quality of the work become harder to separate from the person’s
understanding?” Take one brief example; return to it during the
claims.
Third minute: Show the
five claims. Give people 20 seconds
to read. Ask which they would challenge, take two nominations, and
select two claims. If no one starts, use claims 1 and 4.
Clarify if needed: By an agent, we mean AI that can
carry out several steps toward a goal. An agent might find evidence,
draft a plan, and send it to someone when authorized. Judgment can
be delegated at any of those steps.
3–10 minutes · First claim
Find the judgment inside a real example.
Ask: Read the selected claim and its main question.
Invite one concrete example before offering the example on the site.
-
About two minutes: hear the example. “What did the person ask AI
to do?”
-
About three minutes: expose the choice. “Who chose the goal or
standard? What did the output make harder to notice?”
-
About two minutes: invite an objection. “When would this
assistance strengthen judgment? What would distinguish those
situations?”
Transition: “We disagree about ___. What happens
when we try a different example?”
10–17 minutes · Second claim
Compare a second example.
Do: Open the second claim. Invite a different
panelist or participant to offer an example. Ask which judgment
changed hands, then invite a different view.
Push: “Does our first explanation still work here?
What might AI notice that a person would miss? Who can challenge or
redirect the work before it has consequences?”
At minute 17: Summarize the disagreement in one
sentence. Move on even if it is unresolved.
17–22 minutes · Educational implications
Apply it to something you teach or lead.
“Choose something you teach, assess, or lead. Which judgment needs
to stay visible? How would you design the work so someone practices
that judgment while using AI?”
Do: Allow one minute to think or talk with a
neighbor. Use the remaining four minutes for two contrasting
examples and follow-up.
Listen for: Ask participants to name a capability
and describe where they would practice it. They might explain a
choice, then adapt it when the conditions change.
Test: “Could someone meet that requirement without
understanding the decision? How would you make the demonstration
accessible?”
22–25 minutes · Close
Leave with one judgment to make visible.
Show: The
closing question. Give everyone
30 seconds to form an answer. Invite two or three short responses.
“Which choice will you pay closer attention to after today? What
would you ask the person responsible to explain?”
Finish: Reflect back one useful distinction and one
unresolved disagreement. Point to the
reading room and
sources for continued exploration.
Adapt for 20 or 30 minutes
20 minutes: 0–3 opening; 3–13 discuss one claim;
13–17 implications; 17–20 close.
30 minutes: Keep the 25-minute opening and
two-claim discussion through minute 17. Add five minutes at 17–22
to test the school example. Discuss
implications at 22–27 and close at 27–30.
If the start runs late, reduce the number of claims. Preserve the
implications and closing question.