This is really about training AI, isn’t it. They’ve tapped nearly all the sources of human text output already, so now they want to create as much more of it as possible, as quickly as possible. They will tap into conversations and use it as a new data source, mark my words.