wibbit
← All courses·Course 3 of 5·In progress
🎨

How AI Sees and Creates

Pixels, vision, generative AI, and what “real” means now.

The course about images. Kids zoom into pixels, build their own filters, train a tiny image classifier, then watch diffusion models turn noise into art. They finish by reckoning with deepfakes, consent, and provenance.

5
Modules
24
Lessons
~5.5 hours
Length
8 and older
Ages
Courses 1 & 2
Prereq
The big idea

"AI sees numbers. We taught it to make pictures."

That's the throughline. Every module reinforces it from a different angle — and every lesson ends with the kid being able to demonstrate it.

Source material
Built on the C2PA, Distill.pub, and the Karras et al. StyleGAN papers
We translate research-grade ideas into something a curious kid can play with.
The full curriculum

5 modules. 24 lessons.

Each module ends with a six-question challenge. Pass five, earn the badge.

🔍
Module 1
A New Way of Seeing
Pixels, filters, edges

AI doesn’t see pictures. It sees grids of numbers — 0 to 255 per pixel, three times over. Once you see what it sees, vision AI gets a lot less mysterious.

👁️
Earn the badge
Pixel Detective
Lessons in this module
  1. 1Zoom in until pixels appear
  2. 2Every picture is a number grid
  3. 3Build a filter with 9 numbers
  4. 4Edge detectors — what AI learns first
  5. 5Compare your filter to AI’s
  6. Challenge: Pixel Detective
🏗️
Module 2
Seeing in Layers
CNNs, pooling, classification

Edges become shapes. Shapes become parts. Parts become objects. Each layer of a vision AI sees something deeper than the last.

🧱
Earn the badge
Pattern Stacker
Lessons in this module
  1. 1Ride the elevator through CNN layers
  2. 2Pooling: throw away the boring parts
  3. 3The classifier guesses — and its confidence
  4. 4When the classifier is fooled
  5. 5Why depth matters
  6. Challenge: Pattern Stacker
🎨
Module 3
AI the Artist
Latent space, GANs, diffusion

There’s a hidden map where similar images live near each other. Modern AI navigates that map by learning to undo noise, step by step.

🌈
Earn the badge
Generator
Lessons in this module
  1. 1The idea map (latent space)
  2. 2GANs: Faker vs. Detective
  3. 3Diffusion: from noise to picture
  4. 4Text-to-image with guidance
  5. 5Why some prompts work better
  6. Challenge: Generator
🔗
Module 4
Words and Pictures Together
Multimodal AI

The same attention trick that worked on words works on image patches too. That’s how AI can answer questions about a picture you just took.

🌉
Earn the badge
Bridge Builder
Lessons in this module
  1. 1Slice an image into patch-tokens
  2. 2Vision Transformers attend to patches
  3. 3Shared embedding space (CLIP)
  4. 4Visual question answering
  5. 5When AI hallucinates about images
  6. Challenge: Bridge Builder
🪞
Module 5
Real or Fake?
Deepfakes, consent, provenance

When AI can create anything, what does "real" mean? We learn to spot fakes, ask about consent, and look for the receipts.

🛡️
Earn the badge
Truth Seeker
Lessons in this module
  1. 1Deepfakes: a 10-year timeline
  2. 2Spot the fake (an arms race)
  3. 3Whose data trained the artist?
  4. 4Provenance: showing the receipts
  5. 5Your media literacy toolkit
  6. Challenge: Truth Seeker

Want a heads-up when this lands?

We’re still building this one. Drop your email and we’ll let you know the day it ships — no spam, no other lists.

Join the waitlist →Browse other courses