Paste-ready course blocks.
This operational section keeps the literacy-first premise inside the actual ij8 capability envelope: cohort-level capability grants, per-lesson media-type constraints, one opening prompt per lesson, aiGuidance items, timers, and advisory scoring. The tutor should nudge understanding, prediction, and reflection — not just output quality.
Course config
Title: TMP AI Boot Camp 2026
Description: AI literacy through creative practice. Four days of hands-on experiments in image, video, 3D, audio, and creative code help high school age students at any starting point understand how generative systems behave, how computation works, and how to think with both. The portfolio is the receipt, not the point.
Difficulty band: Easy (Day 1) / Medium (Day 2-3) / mixed (Day 4)
Publish mode: Enrollment (cohort-only)
Cohort: TMP-AI-2026 (1 cohort, ~15 students)
Eight lesson blocks
Lesson 1.1 - First prompts
Lesson 1.1 - First prompts
mediaTypes: ['image']
difficulty: easy
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Welcome. Make five images from one theme. Treat the first output as a draft, then explain what each iteration taught you about how the model responds to language.
aiGuidance:
- If the student keeps the first image, ask what they changed in the prompt and why.
- Ask the student to predict what one changed word will do before they run the next generation.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: do not generate or accept prompts that describe real, identifiable people.
startingCode: null
Lesson 1.2 - Style and specificity
Lesson 1.2 - Style and specificity
mediaTypes: ['image']
difficulty: easy
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Choose one Day 1 image and rebuild it with three specificity layers: subject, composition, and style evidence. Explain which layer gave the most control.
aiGuidance:
- Ask what visual evidence would prove the style constraint worked.
- Push from mood words toward observable details: lens, surface, palette, composition.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: cite any artist or named work used as a reference.
startingCode: null
Lesson 2.1 - Edits and masks
Lesson 2.1 - Edits and masks
mediaTypes: ['image']
difficulty: medium
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Use one image as source material. Produce an inpaint, an outpaint, and a style-transfer edit, then explain what the model preserved and what it invented.
aiGuidance:
- If an edit changes too much, suggest a smaller mask or a more literal instruction.
- Ask the student why the model may have filled the selected region that way.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: do not upload or modify identifiable photos of classmates or minors.
startingCode: null
Lesson 2.2 - Time or dimension
Lesson 2.2 - Time or dimension
mediaTypes: ['video', '3d']
difficulty: medium
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Choose video or 3D. Lift one still into motion or volume, then describe what information the model had to infer beyond the original image.
aiGuidance:
- Ask for one camera move and one subject action, not a paragraph.
- Ask the student to inspect what stayed coherent and what drifted or warped.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: use only objects or self-created imagery with consent for 3D scans or likenesses.
startingCode: null
Lesson 3.1 - Sketch fundamentals
Lesson 3.1 - Sketch fundamentals
mediaTypes: ['code']
difficulty: medium
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Build a p5.js sketch with controllable parameters. Change one parameter or seed at a time and explain what remains deterministic.
aiGuidance:
- When code breaks, ask what changed since the last working version.
- Ask the student to predict what a seed or parameter change will do before running it.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: keep the starting sketch small enough to read aloud.
startingCode: let seed = 327;
let density = 12;
let palette = ['#0a0a0a', '#c01818', '#ffffff'];
function setup() {
createCanvas(720, 720);
noStroke();
}
function draw() {
randomSeed(seed);
background(palette[2]);
for (let x = 40; x < width; x += width / density) {
for (let y = 40; y < height; y += height / density) {
fill(random(palette));
ellipse(x, y, random(8, 38));
}
}
}
Lesson 3.2 - Three dimensions or shaders
Lesson 3.2 - Three dimensions or shaders
mediaTypes: ['code']
difficulty: medium
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Choose three.js or GLSL. Build one small system with two visible controls, then explain how parameters change the behavior of the system.
aiGuidance:
- Ask the student to name the coordinate or uniform they are controlling.
- If the output feels random, ask what rule or parameter could make it legible.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: keep WebGL effects classroom-safe and avoid flashing patterns.
startingCode: null
Lesson 4.1 - Polish and package
Lesson 4.1 - Polish and package
mediaTypes: ['image', 'video', '3d', 'code']
difficulty: mixed
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Select two or three strongest pieces. Write process notes: what you made, how you iterated, what the model contributed, and what you learned.
aiGuidance:
- If the process note only describes the tool, ask for one sentence about revision history.
- Ask the student to label where AI assisted, generated, or transformed the work.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: label work as AI-assisted or AI-generated where appropriate.
startingCode: null
Lesson 4.2 - Showcase and reflection
Lesson 4.2 - Showcase and reflection
mediaTypes: ['image', 'video', '3d', 'code']
difficulty: mixed
masteryThreshold: 0.70
scoringEnabled: true (advisory only - instructor confirms credit; code judge is meaningful mainly for Day 3)
timerMinutes: 50
openingPrompt: |
Prepare for the gallery walk. Share one work and one learning claim: something you now understand about AI or computation that you did not understand on Tuesday.
aiGuidance:
- If share-gallery is unavailable, guide the student to export MP4 or files and attach process notes in the shared folder.
- Ask peer feedback to name one visible choice and one question about how it was made.
- When the student notices something — anything — about how the model behaved, ask them to put it in their own words and write it down. That note is part of the deliverable.
- Require: do not publish identifiable minors or uncited reference-based work.
startingCode: null