Build a Teachable Machine Image Classifier With Your Kid

PromptHacker Issue 15 • July 19, 2023

Children today are surrounded by artificial intelligence, from the algorithms recommending videos on YouTube to the facial recognition unlocking their tablets. However, most kids experience AI as passive consumers rather than active creators. To prepare the next generation for an AI-driven world, we must demystify how these systems work. You do not need a computer science degree or expensive software to teach your child the fundamentals of machine learning. Using a free, web-based tool developed by Google, you and your child (ages 8 to 16) can build, train, and test a custom computer vision model in under twenty minutes. This hands-on project turns abstract technology into a tangible, creative experience, showing them that AI is simply a tool that learns from the data we provide.

The project featured here is a "Focus vs. Distraction" AI Detector . Your child will train a machine learning model to recognize when they are holding a book or notebook (representing "Focus Mode") versus when they are holding a smartphone or video game controller (representing "Distraction Mode"). This project requires nothing more than a computer with a webcam and an internet connection.

To begin, open a web browser and navigate directly to the Google Teachable Machine website at https://teachablemachine.withgoogle.com/ .

Step 1: Create a New Image Project

Once you are on the homepage, click on the "Get Started" button. You will be presented with three options: Image Project, Audio Project, and Pose Project. Select Image Project , and then choose Standard Image Model . This opens the training workspace, which is divided into three clear columns: Classes (where you define the categories), Training (where the computer learns), and Preview (where you test the model).

Step 2: Define and Name Your Classes

In machine learning, a "class" is a category or label that the computer uses to group similar images. By default, Teachable Machine provides two classes named "Class 1" and "Class 2." Have your child rename these classes to make the project intuitive:

• Click the pencil icon next to "Class 1" and rename it to Focus Mode .
• Click the pencil icon next to "Class 2" and rename it to Distraction Mode .

Step 3: Gather the Training Data

Now comes the active building phase. Your child will use the computer's webcam to capture examples of each class. This is called gathering "training data."

• Training Focus Mode: Click the "Webcam" button inside the Focus Mode box. Have your child hold up a book, a notebook, or a pencil in front of the camera. Click and hold the "Hold to Record" button. Encourage your child to move the book slightly, tilt their head, change their distance from the camera, and use different angles while recording. Capture about 100 to 150 images. This variety helps the computer learn what a "book" looks like under different conditions.
• Training Distraction Mode: Click the "Webcam" button inside the Distraction Mode box. Have your child hold up a smartphone, a tablet, or a video game controller. Repeat the recording process, capturing another 100 to 150 images while moving the device around, changing hands, and tilting it.

Step 4: Train the Model

With the training data captured, it is time to train the neural network. In the middle column, click the blue Train Model button.

Important Note: Tell your child to leave the browser tab open and active during this process. It should take less than a minute. Behind the scenes, the browser is analyzing the pixels in the images, finding patterns (like the rectangular shape of a phone versus the larger, textured surface of a book), and building a mathematical model to tell them apart.

Step 5: Test and Play with the AI

Once training is complete, the third column (Preview) will activate, turning on the webcam. Have your child test their creation:

• Hold up the book. The progress bar for "Focus Mode" should shoot up to 90% or 100%.
• Hold up the smartphone. The progress bar for "Distraction Mode" should light up.
• Try holding up something completely different, like a coffee mug or a toy. Ask your child: "Why is the computer confused?" This is a perfect transition into understanding how AI makes decisions based solely on what it has been taught.

If the model is not performing accurately, encourage your child to troubleshoot. This is a vital part of engineering. Have them look at the training images. Are there background objects, like a bright lamp or a family pet, that are confusing the computer? If so, they can delete those images, capture new ones with a cleaner background, and retrain the model. This teaches them that the quality of the output depends entirely on the quality of the input.

Parent and Educator Sidebar: The Core AI Concept

The Core Concept: Supervised Learning

Supervised learning is a type of machine learning where the computer is trained using labeled examples (such as "Focus Mode" and "Distraction Mode"). The computer does not actually understand what a book or a phone is. Instead, it analyzes thousands of pixels to identify mathematical patterns. If the model is only shown blue books during training, it might conclude that all blue objects are books. This is how bias is introduced into real-world AI systems.

Conversation Starters for Your Child:

• What happens if you hold up a blue notebook instead of the book used during training? Does the computer still recognize it? Why or why not?
• If the goal is to train an AI to drive a self-driving car, what kinds of classes and images would be necessary to ensure safety?
• How does facial recognition software on a tablet compare to the model built today?

Action Steps Summary

Inspire the next generation of builders.

Spend twenty minutes this weekend building this project with your child. It is a simple, powerful way to move them from passive technology users to active technology creators.