Activity: The Cat-or-Not Cat Detective Game

Objective:

Help kids understand how AI makes decisions using binary choices (Yes/No, 1/0). The goal is to classify images as either "Cat" or "Not a Cat."

Materials:

  • A set of printed or digital images (a mix of cats, dogs, objects, and other animals).

  • A simple checklist of "Cat Features" (e.g., whiskers, pointy ears, fur, etc.).

  • "Detective Badges" (optional, for fun!).

  • A whiteboard or large sheet of paper to track decisions.

  • Small tokens or stickers (optional, for rewards).

Linear Regression
Linear Regression

Setup:

  1. Introduction:

    • Start by explaining that they will pretend to be AI detectives.

    • Share a quick story: "Imagine you’re training a robot to recognize cats. The robot doesn’t know what a cat is, so you have to teach it step by step!"

  2. Create the Checklist:

    • With the kids, list features that make something a cat. Keep it simple, such as:

      • Whiskers? Yes/No

      • Pointy ears? Yes/No

      • Fur? Yes/No

      • Makes a "Meow" sound? Yes/No

      • Has a tail? Yes/No

    • Each feature represents a "Yes" or "No" decision—a binary choice.

How to Play:

  1. Round 1: The Detective Game:

    • Show an image to the kids (one at a time).

    • Ask the group to go through the checklist:

      • Example: “Does this have whiskers?” Kids shout, "Yes!" or "No!"

    • If the image has most of the “cat features,” classify it as “Cat.” If not, classify it as “Not a Cat.”

  2. Round 2: Training the AI:

    • Pick a few kids to play the "robot" that only answers Yes/No based on the checklist.

    • The rest of the group acts as the AI trainers, giving instructions like, "Does it have pointy ears?"

    • The robot makes decisions until the group agrees if it’s a “Cat” or “Not a Cat.”

  3. Round 3: Introduce Trickier Cases:

    • Add images that are harder to classify, like lions, tigers, or cartoon animals. Discuss why some decisions are harder for both people and AI.

    • Explain that this is why AI needs lots of examples to learn!

Discussion Questions:

  • “What was easy or hard about deciding if something was a cat?”

  • “How does the checklist help us make decisions?”

  • “Do you think AI always gets things right? Why or why not?”

  • “What if a picture of a cat was blurry—how would that change your decision?”

Extensions for Older Kids:

  1. Scoring the AI:

    • Assign points for how many “Cats” the AI robot got right or wrong. Introduce the idea of "accuracy" in AI decision-making.

  2. Teaching AI with New Data:

    • Allow kids to add new features to the checklist, like “small size” or “lives indoors,” to improve the AI's decisions.

Takeaway:

By playing this game, kids learn:

  • How binary decision-making works.

  • Why AI relies on clear rules and examples.

  • That AI gets better with training, just like they do when learning new skills.

And most importantly—they'll have lots of fun being Cat Detectives! 🐾