Program explainer
What is machine learning?
Normal programs are told the rule. A machine learning model is shown examples and works the rule out itself. That difference is the whole subject — and you can do it below in about twenty seconds.
Teach it yourself
How big
Nobody wrote the rule "big and loud means dog". The model drew that line from your examples alone — and if your examples are one-sided, the line comes out wrong. That last sentence is most of what makes AI hard, and we teach it deliberately.
What they'll actually make
Models they train
An image recogniser, a sentiment classifier that reads whether a sentence is happy or angry, a face filter, and a voice recogniser — trained on data they collect, with accuracy they can measure.
Python underneath
The visual tools come first, then the code does. Students move to Python and write the loops, conditions, and data handling that the models sit on top of.
Applications, not demos
Models get wired into working front-ends, so the final project is something a person can use rather than an accuracy score in a notebook.
Why choose this program
Built, not consumed
Most children only ever use AI. Training a model — collecting the data, testing it, watching it fail — is what turns it from magic into something they know the shape of.
Bias is taught, not skipped
Students see their own models go wrong because of the examples they chose. It's the most direct ethics lesson available, and it lands because they caused it.
Real Python
Not a toy language. Python is what the field actually runs on, and what your child would use in a university course or a first job.
Measurable progress
Model accuracy is a number. You can see it improve between lessons, which makes this one of the easiest programs to follow from the outside.
The tools: Teachable Machine and Machine Learning for Kids to start, then Python — all free.