I would argue this particular point. Most of what I do for a living as an ML engineer/researcher is *not* statistical analysis. There's plenty of both in the mix, but they are not the same. If you were to call it "statistical modeling", I might be more inclined to agree with you.