LittleAIMaster Activity Pack
Machine Learning for Families
mixed ages, roughly 4–16 · 3 activities · about 30 minutes each
Print this and work through one activity at a time. Each one ends with something you can check, so you can tell whether the idea actually landed rather than guessing.
1. Sorting game with no computer
Ages 3–7 · about 15 minutes
You need: A pile of buttons, blocks, or socks
- Sort the pile into two groups without saying the rule out loud.
- Let the child guess the rule you used.
- Swap roles — the child sorts, you guess.
- Add one object that does not fit either group and talk about what to do with it.
What went wrong / what surprised us:
2. Train a rock-paper-scissors classifier
Ages 7–14 · about 30 minutes
You need: A laptop with a webcam, Teachable Machine
- Create three classes: rock, paper, scissors.
- Record about thirty webcam samples of each hand shape.
- Train the model and test it live.
- Now test it with the child's other hand, or in a darker corner of the room, and note where it breaks.
- Add samples covering those failures and retrain.
What went wrong / what surprised us:
3. Deliberately bias a model, then fix it
Ages 10–18 · about 45 minutes
You need: A laptop with a webcam, Teachable Machine
- Train a two-class model using samples of only one person.
- Test it on a second person and record how much accuracy drops.
- Write down a prediction for why before changing anything.
- Retrain including the second person and measure the difference.
What went wrong / what surprised us:
A note on tools: everything recommended for mixed ages, roughly 4–16 is free and runs in a browser. If a product asks you to pay before your child can finish an activity, there is almost always a free equivalent.
littleaimaster.com · Machine Learning for Families
Want the next twenty activities?
The app turns these one-off sheets into a sequence with progress you can actually see. Free to start.
Or get the mobile app