Build a Teachable Machine Image Classifier With Your Kid

Demystify artificial intelligence by training a custom computer vision model in your browser in fifteen minutes.

Most children today interact with artificial intelligence constantly, whether through video recommendation algorithms, voice assistants, or school research tools. However, to many young minds, AI feels like magic (an invisible, all-knowing force that simply works). To prepare the next generation of leaders, we must demystify this technology. The best way to show them how AI works is to let them build and train a model themselves.

Using a free tool created by Google, you and your child can build a fully functioning computer vision model in less than twenty minutes. There are no paid subscriptions, no coding requirements, and no specialized hardware setups needed. All you need is a laptop or desktop computer with a standard webcam and a few household objects.

Getting Started with Teachable Machine

To begin this project, sit down with your child at a computer and navigate to https://teachablemachine.withgoogle.com/ . This platform allows users to train machine learning models to recognize images, sounds, or poses directly in the web browser. The data stays local to your browser, making it a secure and private environment for learning.

Click on the Get Started button, then select Image Project , and choose the Standard Image Model . You will be presented with a clean interface showing two empty classes (categories) and a training panel. Explain to your child that they are about to act as the teacher, and the computer is the student.

Step 1: Define Your Classes

First, decide what you want your model to recognize. A classic, engaging setup is classifying two different household items or hand gestures. For example, you can train the model to distinguish between a "Mug" and a "Book", or between a "Thumbs Up" and a "Thumbs Down" gesture.

Rename Class 1 to your first object (e.g., "Mug") and Class 2 to your second object (e.g., "Book") by clicking the pencil icon next to the class names.

Step 2: Gather Training Data

This is where the active building begins. Click the Webcam button under your first class. Hold your first object (the mug) up to the camera.

Instruct your child to click and hold the Hold to Record button while slowly moving the object around. They should rotate the mug, move it closer to the camera, pull it further away, and tilt it. Explain that the computer needs to see the object from many different angles to truly understand what it is. Aim to capture at least 100 to 150 image samples.

Repeat this exact process for the second class (the book), capturing another 100 to 150 images of the second object from various angles and distances.

Step 3: Train the Model

Once both classes have sufficient image samples, click the blue Train Model button in the middle column.

It is crucial to tell your child not to close or switch browser tabs during this process. The computer is currently analyzing all the pixels in the images, looking for patterns, shapes, and colors that distinguish the mug from the book. This training process typically takes thirty seconds to a minute.

Step 4: Test and Iterate

Once training is complete, the Preview panel on the right will activate, showing a live feed from your webcam.

Have your child hold up the mug. The output bar at the bottom should instantly light up, showing 100% confidence for "Mug". Now, have them hold up the book and watch the bar switch to "Book".

To make this a true learning experience, try to break the model. What happens if you hold up a different mug that was not in the training data? What happens if you hold up a hand instead of a book? If the model makes a mistake, explain that this is called a false positive. Show your child how to fix it by adding more diverse training images (for example, adding images of different types of mugs to the first class) and retraining the model.

Parent & Educator Sidebar

Core AI Concept: Supervised Learning This project demonstrates supervised learning, which is the process of training an AI model using labeled data. The computer does not inherently know what a mug or a book is. It simply looks for mathematical patterns in the pixels of the images we labeled. If we feed it biased or limited data (for example, only red mugs), it will struggle to recognize a blue mug.

Conversation Starters:

• "How do you think a self-driving car uses this exact technology to recognize stop signs versus speed limit signs?"
• "If we only trained our model with our faces in the background, do you think it would work if we took the computer outside? Why or why not?"
• "Why is it important for the people who build AI to use many different types of examples when training their models?"

Action Steps Summary

• Go to https://teachablemachine.withgoogle.com/ on a computer with a webcam.
• Create an Image Project and define two distinct classes using household items.
• Record 100 to 150 webcam images for each class, capturing different angles and distances.
• Train the model and test its accuracy in real-time using the preview window.
• Use the sidebar questions to discuss how training data impacts real-world AI systems.

Looking for more practical ways to introduce your family or team to AI?