If you're looking at a way to teach a basic AI for beginners (yes, even teens), you could just try Markov chains. You can make an easy garbage text generator

If you're looking at a way to teach a basic AI for beginners (yes, even teens), you could just try Markov chains. You can make an easy garbage text generator / automatic text-completion (à la text messages) with markov chains, they are quite intuitive to understand and easy to implement. You can even easily make an artificial player that beats you at rock-scissors-paper, in less than 100 lines of code, using only maps and the good old random function. Teens are both surprised, proud and angry when their own code beats them, especially at a game that looks so random ("hey! my own code can detect patterns I'm not even aware of in my playing!") Although that is not for beginners (it is using go), here is a good explanation of what I mean: https://golang.org/doc/codewalk/markov/ https://golang.org/doc/codewalk/markov/. Naive Bayes (and how it can detect spams) is a good second step: it is still rather intuitive and is a real, industry-tested ML algorithm.