Build a Family Image Classifier With Google Teachable Machine

A no-code weekend project that teaches examples, labels, testing, and model mistakes.

Google Teachable Machine lets kids train a computer to recognize images, sounds, or poses without writing code. That makes it perfect for a weekend project where the child can see how examples and labels shape AI behavior.

The goal is not to build a perfect model. The goal is to teach the child that AI learns from examples, that bad examples create bad results, and that testing matters.

Build the classifier project

Train a simple image classifier on three household categories.

• Mug.
• Book.
• Toy car.

Train and test the model

• Open Teachable Machine.
• Choose Image Project.
• Create three classes.
• Add 20 to 30 examples per class using a webcam or uploaded images.
• Train the model.
• Test with new objects, different lighting, and different backgrounds.
• Keep a scorecard of where the model gets confused.

Parent questions for AI literacy

• Which class was easiest for the model?
• Which class was hardest?
• Did lighting change the results?
• Did the background matter?
• What examples would make the model better?

What kids learn about models

They learn supervised learning without needing the term first. They see labels, examples, confidence, testing, and failure modes. They also learn a healthy AI habit: do not just trust the output. Test it.

Source: Google Teachable Machine