This statement is vague, but I do not believe it is true that most published authors are using AI in their work beyond things like grammar checks or research. Some do, sure. But I talk to published authors almost every day and very few of them use LLMs in their work. Why would they? People love to assume new technology must be an improvement for all possible uses. That's rarely the case. AI is great at pumping out mediocre text in large quantities. That makes it excellent for certain things (flooding the KU store, spam emails, clickbait web copy, etc.). Novelists and publishers rarely need more quantity. Quite the opposite, normally. In my experiments, LLMs are mostly poor for improving quality and often detrimental. There’s a reason the conventional wisdom pre-LLMs was authors should log off, avoid distractions, and immerse themselves in the work. thewalrus.ca/new-york-t…