Offline AI learning for kids

MABNI introduces AI through child-friendly activities. Students label examples, train simple models, test predictions, and discuss why human verification matters. The aim is not hype; it is clear classroom understanding of data, labels, prediction, mistakes, and responsible AI use.

Included AI activities

• Cedar Guardians: teach Pip with real reference photos, then explore six robot expeditions powered by your labels.
• AI Shape Trainer: playful offline game about labels, features, and predictions.
• AI Maker: train image models locally, test new photos, and send predictions to a micro:bit robot over USB.
• Teacher dashboard: local-first classroom support and offline progress foundation.