In general, the AI chapter is what I am currently working on most. I mean not "only" on the actual content, but on the applications I want to present in this chapter.
Anton Kochkov has recently motivated this chapter via a dedicated issue for AI/ML applications of Prolog, and I invite you to track its progress at:
For your machine learning section, you might want to consider discussing Metagol [1], a modern ILP system that can learn recursive theories and perform predicate invention.
The core Metagol implementation is tiny and relies on a meta-interpreter (the technique is called Meta-Interpretive Learning). It is a radically different approach to previous ILP systems but much simpler and easier to understand. It should be interesting to anyone who wants to do machine learning with Prolog.
____________
I have added Metagol to the AI chapter, please have a look.