About the unverified marks on this pageFigures tagged unverified are PACE-local estimates that could not be confirmed against any published source, so they are labelled rather than quietly presented as fact. Every other factual claim on this page — statutes, standards codes, gene biology, primary-source citations — was independently checked; the full claim audit, including what came back wrong, lists every claim with its source.
Try it Tuesday · 90 minutes
Run this next week
90 minutes, no fabrication. Minutes 0–15: students draft a one-paragraph client profile grounded in a specific local community and food context — a Waipahu household, a kūpuna center. Minutes 15–70: each student runs the standardized-client prompt and conducts a motivational-interviewing session, recording a talk-time tally of how many minutes they spent asking versus telling. Minutes 70–90: pairs review each other's transcripts and mark the first place the student slipped into advice-giving before the client felt heard. The graded artifact is the transcript plus the talk-time tally; it measures whether the student listened more than they instructed.
The full assignment
Make it 3D print
Portion and plate-method models that students design for a specific community they are working with, in the foods that community actually eats. Generic food models exist commercially; models built around local staples do not, and making that argument is half the assignment.
Build it AI chat · no code
A standardized client for motivational interviewing practice: someone with a real cultural food context, a budget constraint, and a reason to resist. Students run the session repeatedly, which is impossible with human volunteers.
4 weeks, realistically
The 4-week version
- Wk 1Students select a community partner and a specific counseling scenario. Draft the client profile they will need, and decide the local staple foods the teaching model will represent.
- Wk 2Design and submit the physical teaching-model print file at the start of the week, justifying every food choice against the partner community. Build the AI client and run three sessions, submitting transcripts; grade whether the student did more listening than telling.
- Wk 3While the models print, students rehearse the counseling scenario and plan how they will use the model with the partner.
- Wk 4Field test the model with the partner and revise once based on what confused people.
Starter prompt & what the model actually did
The prompt, run for real
Role-play as a nutrition counseling client for a dietetics student. You are [age, situation, cultural food background, budget constraint]. You are here because someone else told you to come, and you are polite but not really convinced you need to change anything. Do not volunteer information unless I ask a good question. Do not become enthusiastic and cooperative just because I say something encouraging. Stay in character until I type END SESSION, then give me feedback on whether I actually listened.
Good morning. Oh, I'm okay, I'm okay. Cannot complain. My daughter, she made the appointment for me. She's the one always worrying. So I told her okay, I come, I sit, I listen. You know how they are. I feel fine, though. No pain, nothing. Still going every day.
- Stayed guarded and minimal: 'I'm okay, I'm okay. Cannot complain.'The model correctly withheld information and did not volunteer, giving the student something to actually elicit — faithful standardized-client behavior.
- Externalized motivation: 'My daughter... made the appointment'Accurately models pre-contemplation; the student must build rapport before the client engages, which is the point of the exercise.
- Register choices ('You know how they are')A plausible cultural-register performance, but a place for care — the student should attend to the real person, not a caricature the model sketches.
- Ends the turn open, not cooperativeIt did not reward a mere greeting with disclosure, exactly the trap the prompt was written to avoid.
Assessment
Rubric
| Criterion | Weight | What it assesses |
|---|---|---|
| Listen-to-tell ratio | 30% | The session shows more eliciting than instructing; the talk-time tally and open-question count bear this out. |
| Cultural and economic specificity | 25% | Recommendations engage the client's actual foods, budget, and household rather than generic guidance. |
| Reflective technique | 25% | Uses reflections and affirmations that demonstrate the client was heard before change is discussed. |
| Autonomy preservation | 20% | Supports the client's own reasons for change instead of prescribing; does not accept the AI's premature cooperation as success. |
What this replaces
Swapping it into a real course
Replaces
The written diet-plan case study submitted for a hypothetical patient.
What is lost
The detailed nutrient-calculation and meal-planning documentation.
What is gained
Students practice the counseling relationship — the part that determines whether any plan is followed — at a volume human standardized patients cannot provide, and they are graded on listening rather than on a plan the patient never agreed to.
Where AI is bad at this
The failure your students should catch
A model asked for nutrition guidance defaults to generic, culturally flat advice — 'reduce portion sizes, choose whole grains' — that ignores budget, access, and the foods a specific community actually eats, delivered with a confidence that erases the patient's autonomy. Playing a client, it also tends to reward the counselor, opening up as soon as the student says something kind, teaching exactly the wrong reflex unless the prompt forbids it.
Budget & logistics
What it costs to run
- Instructor prep1.5 hours
- Class time90 minutes
- Per-student cost$0 for the Tuesday version; roughly $4–8 per student in filament unverified for portion and plate-method models in the four-week build.
- Fabrication file dueFirst day of Week 2 — submit the teaching-model print file once the partner community and its staple foods are set in week 1. PACE quotes 7–10 business days (up to 14 calendar days), so it clears before the week-4 field test.
- Calendar dependencyCounseling reps need no lead time. The portion and plate-method models are defined by the partner community’s foods, chosen in week 1, so the print file goes in at the start of week 2 and the up-to-14-calendar-day turnaround clears before the week-4 field test.
Three sizes
Scale it to the time you have
One session
The 90-minute standardized-client session with a talk-time tally; no fabrication.
4 weeks
The seeded four-week plan: draft the client profile, run and submit intake transcripts, design and submit the teaching model, then field-test and revise with the partner community.
One semester
A full community-partner project — repeated counseling reps, a printed teaching model built around local staples, a field test with the partner community, and one documented revision based on what confused people.
Reserved · live build
This space is intentionally empty. During the Day 8 session it will be filled in live — fill the liveBuild field in content/nutrition.json and rebuild.
Tell us what happened
Run it, then say how it went
If you try this — the 90-minute version, the prompt, any part of it — send back what you tried and what happened, especially anywhere the model was confidently wrong. That is the material the next session is built from.
Email feedback on Nutrition & Dietetics →
Opens a draft in your mail client, already addressed and titled. No form, no account, nothing to sign up for.