Teach the AI to sort objects into categories by giving it training examples
Drag each item into the correct category. The AI is watching and learning from your examples!
A classifier is one of the most fundamental types of AI system. Its job is to look at an input (like an image, a piece of text, or a set of measurements) and assign it to a category. Email spam filters classify messages as "spam" or "not spam." Photo apps classify images by content — "beach," "birthday party," "pet." Medical AI classifies X-rays as "normal" or "abnormal." Classification is at the heart of countless AI applications.
The most important concept in this activity is that AI only knows what you teach it. If you train a classifier with many examples of dogs but no examples of cats, the AI will not know what a cat is. If you train it with only red cars, it might not recognize a blue truck as a vehicle. The quality, quantity, and diversity of training data directly determines how well an AI system performs.
This principle has real-world consequences. Facial recognition systems trained primarily on lighter-skinned faces have been shown to perform poorly on darker-skinned faces. Language models trained on internet text can absorb and reproduce societal biases. Understanding the connection between training data and AI behavior is essential for anyone who will use, build, or be affected by AI systems — which is everyone.
This activity demonstrates supervised learning, where the AI is given labeled examples (you label each item as "animal" or "vehicle") and learns to predict the correct label for new items. Supervised learning is the most common type of machine learning and powers applications from voice assistants to recommendation systems to self-driving cars.
Last reviewed: April 2026 · Aligned with AI4K12 Big Idea 3: Learning