Emotion Classifier with Teachable Machine
Can a machine truly "understand" a smile, or does it just recognize a curve of pixels? Emotion Classifier is a hands-on project that uses image classification to bridge the gap between technical AI literacy and the lived human experience.
The Human Reflection
While we categorize emotions into digital "buckets" (Happy, Sad, Angry), this project interrogates what is lost in translation. I lead students to reflect on the meaning of emotion: Is it a universal biological signal, or a culturally specific performance? By building their own classifiers, learners confront the tension between the fluid nature of human feeling and the rigid logic of data.
AI Ethics in Practice
· The accompanying Teaching Plan moves beyond the "how-to" to address critical ethical frontiers:
· The Myth of Objectivity: How developer bias shapes "universal" emotion sets.
· Emotion Surveillance: The risks of AI inferring internal states in public spaces.
View the full Teaching Plan below to explore how algorithmic skepticism can be turned into an interactive learning journey.