Build a Teachable Machine Image Classifier With Your Kid

Spend twenty minutes building a functional, custom computer vision model with your child using Google's free web tool.

As an executive, you understand that artificial intelligence is rapidly reshaping every industry. However, explaining neural networks and machine learning to your children can feel abstract and dry. Instead of lecturing them, you can spend twenty minutes building a functional, custom image classifier together. Using Google's free Teachable Machine tool, your child will train an AI model to recognize objects, hand gestures, or family pets using a standard computer webcam. This hands-on project demystifies how computer vision works and introduces core machine learning concepts without writing a single line of code.

Why Hands-On AI Education Matters

Children today are surrounded by AI, from video game recommendation engines to facial recognition on tablets. However, they often remain passive consumers of these technologies. To prepare them for a future driven by automation, we must shift their perspective from consumption to creation. When children build and train their own models, they realize that AI is not magic: it is a tool that relies on human input, data, and pattern recognition.

This activity is designed for children aged 8 to 16. It requires no paid subscriptions, no specialized hardware, and no prior programming experience. All you need is a laptop or desktop computer with a working webcam and an internet connection.

Step 1: Setting Up Your Workspace

To begin, sit down with your child and open a web browser on your computer. Navigate directly to https://teachablemachine.withgoogle.com/ . Click on the blue "Get Started" button on the homepage. You will be presented with three project options: Image Project, Audio Project, and Pose Project. Select "Image Project" and then click on "Standard Image Model". This opens the training interface, which is divided into three clear columns: Classes, Training, and Preview.

Step 2: Defining Your Classes

In machine learning, a "class" is a category or label that the computer learns to recognize. For this project, we will build a hand gesture classifier. Rename "Class 1" to "Thumbs Up" and "Class 2" to "Peace Sign". You can add a third class by clicking "Add a class" and naming it "Open Palm".

Explain to your child that they are setting up the rules of the game. The computer will try to sort whatever it sees through the webcam into one of these three categories.

Step 3: Gathering Training Data

This is the most interactive phase of the project. Click the "Webcam" button inside the "Thumbs Up" class box. Have your child sit in front of the camera and hold up a thumbs-up gesture. Click and hold the "Record" button to capture images.

Instruct your child to move their hand slightly, tilt it, bring it closer to the camera, and move it further away while you record. This variation is crucial. Explain that if the computer only sees a hand in one exact position, it will struggle to recognize the gesture in different contexts. Capture about 100 to 150 images. Repeat this exact process for the "Peace Sign" and "Open Palm" classes, ensuring you capture a similar number of images for each.

Step 4: Training Your Model

Once you have gathered data for all three classes, move to the middle column and click the "Train Model" button. This process takes about ten to thirty seconds. Make sure to keep the browser tab open and active during this time.

While the progress bar fills, explain to your child that the computer is analyzing the images. It is looking for patterns, such as the shape of the fingers, the angles of the hand, and the contrast against the background. The computer is building a mathematical formula to distinguish between the three gestures.

Step 5: Testing and Breaking the Model

Once training is complete, the "Preview" column on the right will activate your webcam. Have your child make the gestures in front of the camera. They will see real-time probability bars at the bottom of the preview window showing how confident the AI is in its classification.

Now, challenge your child to "break" the model. What happens if they use their left hand instead of their right hand? What happens if you, the parent, try the gestures? What happens if they hold up a toy or a mug?

If the model misclassifies your hand or a different object, explain that this is called "bias" or "overfitting". The model only knows what it has been taught. If it was only trained on your child's right hand, it might not understand an adult's hand or a left hand. This is a perfect transition into discussing how real-world AI systems can make mistakes when their training data is limited.

Parent and Educator Sidebar

Core AI Concept: Supervised Learning

Supervised learning is a type of machine learning where the model is trained on labeled data. In this project, the labeled data consisted of the images categorized under "Thumbs Up", "Peace Sign", and "Open Palm". The computer learns by comparing new inputs to these labeled examples.

Conversation Starters:

• "How do you think a self-driving car uses this technology to see stop signs versus pedestrians?"
• "What would happen if we only trained our model with pictures of your hand, and then a friend tried to use it?"
• "How could we make our model more accurate without changing any computer code?"

Action Steps Summary

• Navigate: Go to https://teachablemachine.withgoogle.com/ and start an Image Project.
• Define: Create classes for "Thumbs Up", "Peace Sign", and "Open Palm".
• Record: Capture 100 to 150 webcam images for each class with varying angles.
• Train: Click "Train Model" and observe the pattern recognition process.
• Test: Use the live preview to test and intentionally break the model.

Save this link for a weekend learning session with your child.