There's been HN threads on it before, but just in case anyone hasn't seen it and wants to dive down the rabbit hole that is biomimetic / natural systems influenced generative work, check out Dan Shiffman's awesome http://natureofcode.com/.
I would recommend 'Generative Design'. http://www.generative-gestaltung.de/ (There is an English edition)
A really beautifully designed book. Great to leaf through for inspiration, but also has decent example code for a variety of techniques.
Processing is still cool, and there is a JavaScript version available and lots of books and online examples. It really kicked it off for me [2].
The term 'Creative Coding' seems to be the one that is sticking and encompassing the widely different approaches to creating art with code and computers and other electronic interfaces and devices. Very exciting time for this field right now. Livecoding is another term you'll hear for creating mostly music live in code, with some still doing visuals [3]. Fluxus was a favorite of mine [4].
openFrameworks, a C++ toolkit has a lot of libraries, including ML and AI stuff, to try out a bunch of different things [5].
There are examples of people using GA (Genetic Algorithms), ANNs (Artificial Neural Networks), GP (Genetic Programming), MCMC (Markov Chain Monte Carlo) methods and more to create artworks, visual, audio or mixed.
The Nature of Code and some YouTube Videos for SuperCollider and other creative coding software touch upon some of this, but you need to Google for specific examples, since people are using them all over the place. My first exposure was a Java Applet where people would look at a generated artwork and hit +1, 0, -1, for like, neutral, dislike, and the applet purportedly used a genetic algorithm to make a new piece. The piece would evolve with the user being the fitness function by participating in judging the pieces.
I am playing with Darknet, a C-based CNN (Convolutional Neural Network) that can utilize the GPU, and also work with OpenCV [6]. You can use it to generate text based upon famous writer samples, or recognize objects in an image or video.
[1] extempore.moso.com.au
[3] toplap.org
As well as using a pretty interesting fractal algorithm for the content it uses a genetic algorithm for filtering of results and progression. He posts everything to GitHub too if you're interested: https://github.com/scottdraves/electricsheep https://github.com/scottdraves/flam3