👩‍🏫 Teacher Overview

• Grade Band: Elementary (3–5), Middle (6–8)
• Objective: Students understand that AI learns patterns from examples, and sometimes misclassifies new data.
• Time Needed: 30–40 minutes
• ISTE Standards:
• Knowledge Constructor: Students critically curate resources to construct knowledge.
• Computational Thinker: Students break down problems and recognize patterns.
• Key Vocabulary: AI, pattern, classifier, training data, prediction, mistake.

📦 Materials

• Printed image cards (dogs, cats, foxes, raccoons, simple shapes)
• “Feature” cards (ears, tails, paws, whiskers, fur patterns)
• Sorting mats (Yes Dog / Not Dog)
• Student worksheet (included below)
• Optional projector/slides for whole-class demo

📝 Teacher Steps

1. Hook (5 min)

• Show 3–4 dog images. Ask: “What makes this a dog?” (students list features).
• Record features on board: paws, ears, tail, fur, nose.

2. Training Data Game (10 min)

• Hand out dog image cards (all training set = dogs).
• Students sort into “Dog” pile — noticing repeating features.
• Teacher reinforces: “The AI is learning: tails, floppy ears, paws = dog.”

3. New Data Challenge (10–15 min)

• Give mixed cards: cats, foxes, raccoons.
• Students must “classify” → Dog or Not Dog.
• Intentionally include tricky examples (e.g., a cat with floppy ears).
• After sorting, reveal which are actually not dogs.

4. Discussion & Reflection (10 min)

• Why did you think the raccoon was a dog?
• How do mistakes happen in AI?
• Where would it matter if AI made a mistake (e.g., self-driving car, medical diagnosis)?

🧒 Student Worksheet

Name: _____________________

Date: _____________________

Part 1 – Training the AI

Look at the pictures of dogs. Circle the patterns you see:

• 🐾 Paws
• 👂 Ears
• 🐕 Tail
• 👃 Nose
• 🎨 Fur patterns

Draw one more feature you notice: ____________________

Part 2 – Test the AI

For each picture below, decide: Dog 🐶 or Not Dog 🚫

Check the box:

Part 3 – Reflection

1. Did the “AI” (you!) ever make a mistake? ___ Yes / No
2. Why do you think mistakes happen?
3. How could AI do better?
1. Learn about more types of dogs? with and without tails, fur etc.
2. Learn about cats, foxes, raccoons as well?
3. Learn about how they sound?

The more AI learns, the better it can get, but there can always be mistakes.

🎨 Image Set

For the printable pack, we’ll want:

• Training set (Dogs only):
• 3 clear dog images (different breeds)
• Test set (Mix):
• 2 cats, 1 fox, 1 raccoon, 1 tricky dog (e.g., corgi without a tail)
• Feature cards: simple icons of paw, ear, tail, whiskers

Dachshund has paws, tail, ears, fur, whiskers

English Sheep dog has paws, tail, ears, fur, whiskers

Golden retriever has paws, tail, ears, fur, whiskers

Corgi has paws, ears, fur, whiskers but no tail

Dalmatian has paws, tail, ears, fur, whiskers

Bengal cat has paws, tail, ears, fur, whiskers

Persian cat has paws, tail, ears, fur, whiskers

Fox has paws, ears, tails, whiskers

Raccoons have ears, tails, whiskers, but hands instead of paws

At the end of the activity students will have realized that it’s possible for AI to make mistakes, and can start to understand how and why these mistakes can occur.