Build a Teachable Machine Image Classifier With Your Kid

PromptHacker Issue 5 Companion Article

In an era where artificial intelligence is reshaping every industry, teaching children how these systems work is a vital modern skill. However, reading about algorithms can be dry and disengaging for young minds. The most effective way to demystify artificial intelligence is through hands-on creation. By building a custom image classifier, children ages 8 to 16 can experience the core mechanics of machine learning firsthand. Using Google's free Teachable Machine platform, you and your child can train a computer vision model in less than twenty minutes using nothing more than a standard laptop or tablet webcam. This project requires no coding experience, no paid subscriptions, and no specialized hardware, making it an ideal weekend activity for busy executives to connect with their children while exploring cutting-edge technology.

Most children interact with artificial intelligence daily through video game recommendations, facial recognition on tablets, or voice assistants. These interactions are passive, often leaving kids with the impression that AI is a magical, all-knowing entity. By building an image classifier, children shift from passive consumers to active creators. They learn that artificial intelligence is simply a tool that recognizes patterns based on the data humans provide. This realization builds technological confidence and critical thinking skills that will serve them throughout their academic and professional lives.

Step 1: Access the Platform and Choose Your Project

To begin this project, clear a small workspace on a desk or table. You will need a computer, laptop, or tablet with a working webcam and an internet connection. Open a web browser and navigate directly to the official Google Teachable Machine website at https://teachablemachine.withgoogle.com/ to start the project.

Click on the Get Started button on the homepage. You will be presented with three project options: Image Project, Audio Project, and Pose Project. For this activity, select the Image Project option, and then choose the Standard Image Model. This setup allows the webcam to capture visual data and train the system to recognize different objects or gestures.

Step 2: Define Your Classes and Collect Training Data

Explain to your child that the computer needs to learn what different objects look like. To do this, you will create distinct categories called classes. For this guide, we will build a classifier that distinguishes between a coffee mug and a pen, two items easily found on any executive's desk.

• Rename Class 1 to Mug by clicking the pencil icon next to the label.
• Click the Webcam button under the Mug class. Hold your coffee mug in front of the camera.
• Click and hold the Hold to Record button. Move the mug slightly, tilting it, bringing it closer to the camera, and moving it further away. Capture at least 100 image samples. This variety helps the model learn what a mug looks like from different angles.
• Rename Class 2 to Pen.
• Hold a pen in front of the webcam and repeat the recording process, capturing another 100 image samples from various angles.

Step 3: Train Your Machine Learning Model

Once you have collected the training data for both classes, look at the middle column labeled Training. Click the blue Train Model button. Instruct your child to leave the browser tab open and avoid closing the laptop while the model trains. This process usually takes about ten to thirty seconds. During this time, the browser is processing the images, extracting key features, and building a mathematical model to distinguish between the mug and the pen.

Step 4: Test and Refine Your Model

Once training is complete, the Preview column on the right side of the screen will activate, showing a live feed from your webcam. Now comes the exciting part for your child: testing the model.

• Hold the coffee mug up to the webcam. Watch the output bars at the bottom of the preview window. The bar labeled Mug should shoot up to 100%, indicating high confidence.
• Hold the pen up to the webcam. The bar labeled Pen should now shoot up to 100%.
• Try holding up a completely different object, like a notebook or your hand. Observe how the model struggles to classify it, often guessing wildly between Mug and Pen. This is a perfect moment to explain that the computer only knows what we have taught it.

If the model makes mistakes, click the record button under the corresponding class to add more diverse images, then retrain the model. This iterative process of testing, identifying errors, and refining data is exactly how professional AI engineers build enterprise-grade systems.

Expanding the Project: Advanced Challenges

Once your child has mastered the basic mug and pen classifier, you can challenge them to build more complex models. This keeps the activity engaging for older children (ages 12 to 16) and introduces deeper computer vision concepts.

• The Rock-Paper-Scissors Game: Create three classes named Rock, Paper, and Scissors. Have your child record hand gestures for each. This requires more precise positioning and teaches the model to recognize complex human hand shapes rather than rigid physical objects.
• Facial Expression Detector: Create classes for Happy, Sad, and Surprised. This challenge demonstrates how facial recognition and emotion detection systems work in modern consumer devices. It also highlights how subtle changes in lighting or facial angles can affect the model's accuracy, prompting a discussion on the importance of diverse training data.

Parent & Educator Sidebar

Core AI Concept: Supervised Learning

This project demonstrates supervised learning, where an algorithm learns from labeled training data (the images of the mug and pen). The computer does not actually know what a mug is; it simply recognizes patterns of pixels, colors, and edges associated with the label we provided.

Conversation Starters:

• What do you think would happen if we trained the model with only blue pens, and then showed it a red pen? Would it still know it is a pen?
• How does a self-driving car use this kind of technology to stay safe on the road? What classes of objects does it need to recognize?
• Why is it important for the humans who train AI to use many different types of examples instead of just one?

Action Steps Summary

• Visit the Site: Go to https://teachablemachine.withgoogle.com/ and start a Standard Image Model project.
• Capture Samples: Record at least 100 webcam images of two different household items for Class 1 and Class 2.
• Train the Model: Click Train Model and wait for the browser to process the visual patterns.
• Test and Iterate: Challenge the model with different angles, lighting, or new objects, and add more training data to improve accuracy.

Want more practical AI workflows for executive performance?