Taking that first dive into a new field of study can be daunting. Having to combine multiple fields in order to understand a general, parent field, can quite difficult. Sadly, this is the exact problem most prospective newcomers in the field of AI face. From math to statistics to coding to data analysis - you name it. We're only human. Unless you're absolutely brilliant, it's going to be difficult, if not impossible, to become "master of all."
Artificial Intelligence is a domain of Computer Science - and to understand it you need to understand the basics of how computers work. And what better than to start at a school like Harvard University. (visit CS50 on edx.org). It'll also help you learn programming, which will always be at the core of your machine learning journey. MIT’s Programming for Everybody is a 5-course specialization that'll go in depth on Python (the go-to language for A.I). Also, download anaconda’s distribution of Python 3 - it comes with a plethora of packages and a wealth of useful troubleshooting information. After completing the specialization, you should dive straight into the Applied Data Science with Python specialization. It's more difficult and will require more time.
Next comes the mathematics. There are many places to learn how to understand machine learning math online, and I advise that you see which one suits your learning style! The necessary math skills in Machine Learning include Linear Algebra, Multivariate Calculus, and Statistics. A good place to start is always Khan Academy. If you want something a bit more challenging, Data Science Math Skills is great too. On the other hand if these courses/videos seem too difficult, or you simply don't have enough time, take a look at 3Blue1Brown’s YouTube channel (especially Essence of Calculus and Essence of Linear Algebra playlists).
Andrew Ng, among other things, is one of the most knowledgeable figures in AI. His Machine Learning course is a must-take - it will teach you everything need to know in great detail. If you made it this far, then you can wrap things up with Deep Learning, a 5-course specialization, again by Professor Ng.Links Section :