I suspect that spell-checkers have something to do with it. Or, people just don't pay attention.
But on the correct-grammar/spelling issue more broadly: go and read books by the highly educated intelligentsia of, say, the sixteenth or eighteenth century. Shakespeare, Jonathan Edwards, whoever you like. Read them in original manuscripts, or in editions that have not been "corrected" or standardized by modern editors. What you'll see is that for most of the history of human education and literacy, phonetic spelling has been pretty normal and people have used written grammar that reflects spoken grammar (Jonathan Edwards, for instance, regularly uses "don't" where the "correct" usage is supposed to be "doesn't"). And forget about medieval writers. I have manuscripts that spell the name of THE SAME INDIVIDUAL variously as Odilo, Otilo, Ottilo, Uatilo, and more besides - all by highly educated scribes.
The fact is that the notion of strictly "correct" grammar and spelling independent of actual spoken usage is relatively recent -- late eighteenth/early nineteenth century, if I remember correctly. The point of introducing it was to create a means to distinguish people by class in an age of upward mobility, more widespread education and literacy. Only the very most educated would know the intricacies of the various artificial, arbitrary rules. Of course, democratization in education played along so that every public school made a goal of teaching "correct" grammar and spelling to every child and dismissing them as "uneducated" if they failed to play by the rules.
It's obvious from this that I don't approve of the whole phenomenon. The irony is that, as a teacher, I still use correct usage as one of my grading criteria. The reason is that, for better or worse, "correct usage" is still a litmus test of a person's educational achievement and I can't send my students out into the world with the message that it doesn't matter -- if I don't make them write "correctly," somebody else will penalize them for it.