How should we speak about the processes of Large Language Models and other AI? Language, as humans and other animals use it, is a voluntary and intentional tool which is used to aid interaction. Can we say the same for AI generated text?