Hi all, long time lurker. Question about Machine Learning and data objectives

Been doing online ML courses on my own the last month, I have some statistics experience and ran some significance tests many years ago for my thesis and other activities. As I understand it, ML is a subset of AI. As such, information/historic data is inputted, data split testing and training occur, and, the system provides information that is prediction.

I might not have chosen the best data sample in which to "learn" or predict, but, I used a cloud A.I app in supervised maner and inputted historical government economic information from a CSV data files, mainly economic in nature. Would this be this considered ML per se?

It doesn't seem to be teaching the machine anything other than mathematical suggestion by relation. Nothing more than other statistical estimations I might have made years ago with softare and data

Furthermore can anyone recommend any good data result interpreting books re: ML? I have grasped a general understanding of the algoroitms and how they look to present reliable and valid results but, trying to interpret it as some greater platform than anything I could normally do, and, applying it in plain language(often the tough process which requires repetitive and experience) has eluded me.

Sorry for the lengthy. I hope I wasn't too confusing. I've done a binary test, single regression and KNN, received the results. My ability infer anything more "this test is a valid test result for it", is lacking.

Thanks for your time.