Build a Teachable Machine Image Classifier With Your Kid

A hands-on, zero-coding weekend project to teach children ages 8 to 16 how machine learning models are trained.

Most children today are passive consumers of artificial intelligence. They interact with algorithms through video recommendations, voice assistants, and search engines, but they rarely get to see behind the curtain. To prepare the next generation of leaders, we must shift their perspective from viewing AI as a magical black box to understanding it as a practical tool built on data, patterns, and training.

You do not need a computer science degree or expensive software subscriptions to teach these concepts. Using a free, browser-based tool developed by Google, you and your child can build, train, and test a custom image classifier in less than half an hour. This project is designed for children aged 8 to 16, requiring only a computer with a webcam and a few household objects.

Step 1: Access the Platform

To begin, open a web browser on your computer and navigate to Google Teachable Machine . Click on the "Get Started" button, and select "Image Project" followed by "Standard Image Model". This platform runs entirely in the browser, meaning no data is sent to external servers, and no account creation is required.

Step 2: Define Your Classes and Gather Data

An image classifier works by recognizing patterns in visual data. To demonstrate this, have your child choose two distinct objects from around the house (such as a coffee mug and a pen, or a toy dinosaur and a toy car).

In the Teachable Machine interface, you will see two default categories called "Class 1" and "Class 2". Have your child rename these classes to match the chosen objects. Next, click the webcam button under the first class. Hold the first object up to the camera and click and hold the "Record" button to capture approximately 100 to 150 images. Encourage your child to rotate the object, move it closer and further from the screen, and tilt it to capture different angles. Repeat this exact process for the second object in the second class.

Step 3: Train the Model

Once both classes have sufficient training images, click the "Train Model" button in the middle column. It is crucial to keep the browser tab open and active during this process. The computer is now analyzing the pixels in the captured images, identifying the unique edges, colors, and shapes that distinguish the first object from the second. This process typically takes between thirty seconds and one minute.

Step 4: Test and Break the Model

Once training is complete, the preview panel on the right will activate your webcam. Have your child hold up one of the objects. The model will display a real-time confidence rating (from 0% to 100%) showing which object it thinks it is seeing.

Now comes the most educational part of the exercise: trying to break the model. Have your child hold the object in a different room with different lighting, or cover half of the object with their hand. Observe how the confidence rating drops. This demonstrates that the model does not actually understand what the object is, it only recognizes the specific pixel patterns it was trained on.

Parent and Educator Sidebar

Core AI Concept: Supervised Learning

Supervised learning is the process of training an algorithm using labeled data. The computer does not inherently know what a toy car is. It only learns to associate a specific set of visual inputs with the label we provide.

Conversation Starters:

• "If we only trained the model with a red toy car, why does it struggle to recognize a blue toy car? How does this relate to bias in real-world AI?"
• "How do you think self-driving cars use this technology when they are driving down a busy street?"

Action Steps Summary:

• Open the Tool: Visit Teachable Machine on a laptop or desktop computer with a webcam.
• Gather Training Data: Create two classes and record 100 to 150 webcam images for each household object.
• Train the Model: Click train and wait for the browser to process the visual patterns.
• Experiment and Discuss: Test the model under different conditions and use the sidebar questions to discuss how machine learning works.

Looking for more practical ways to introduce your family to technology?

Make the activity stick

Have the child write down the examples before opening the tool. Three good examples for each label beat twenty rushed examples. If the classifier makes a strange prediction, treat that as the best part of the activity. Ask which examples confused it and what new example would make the boundary clearer.

Use Google Teachable Machine as the starting point because it lets kids see labels, examples, training, and testing in one browser window. No paid account is required for the basic activity, and the child gets a concrete model instead of a vague explanation.

Parent or educator prompt: ask what the model learned, what it failed to understand, and how a biased set of examples could change the result. The core AI concept is training data quality. Kids should leave knowing that models do not magically understand the world, they copy patterns from the examples people give them.