I think English will change in a way that being “wrong” will be part of the language itself.

AI can learn a language, but can’t learn the unpredictability of communities shaped by memes. AI for example doesn’t know memes, because requires processes unable to be emulated via pure logic.