Deep learning utilises multiple layers of "neurons" with some weighted connections to the next layer to convert a large input into a small output. What your script does is uses a probabilistic model to generate semirandom characters based on how often they appear in the input text.

This doesn't mean your script is a bad one - you've clearly understood lists and the essence of functions, so when it comes to Skript I think you at least have an idea of what you're doing, but that doesn't make this a neural network.

It's also quite a strange Markov chain - but from what I've seen I think it does somewhat satisfy the Markov property, at least in the generateText function