Training Data and Bias
by Anonymous
Add and remove animal pictures from the baskets to see how training data shapes—and biases—the AI's answers.
About this app
Training Data and Bias is an interactive AI literacy activity for exploring how examples affect an AI’s answers. Players add or remove cat, dog, and rabbit picture cards from baskets, test the AI with animal images, and observe how missing or uneven data can produce uncertain or biased results. Four missions across three sections guide learners through teaching, correcting imbalance, and identifying the reason behind an AI response.
Use cases
- Elementary students learning AI and machine learning basics
- Teachers introducing training data and data bias
- Families discussing how examples shape AI decisions
- Beginners practicing computational thinking through a visual simulation
Features
- Add or remove cat, dog, and rabbit cards from training baskets
- Test the simulated AI with animal pictures and view its answer
- Complete four teaching missions and four bias-correction missions
- Compare balanced, missing, and heavily skewed training data
- Choose explanations for why the AI answered a certain way
- Optional sound toggle, reset controls, and step-by-step progress