if you alter your choice of words and the way you speak / write because you’re concerned about sounding like chatGPT, or if you police others’ words in the same way, you’re letting a large language model reshape language itself. And I’m not entirely convinced any of us have thought that through.