There definitely seems to be a continuum. I’m legitimately confused about using an LLM to generate actual text, since that seems like the easy part. I am using one to help write a novel, but it’s as a research assistant. For example, I struggled for years using tools like Vulgar to create new languages. But, it turns out LLMs are great at creating them. A list of things I routinely use the LLM for:
Finding the perfect word. Often times, I am trying to describe something and I can’t even come up with a word to look up in the thesaurus. So, “how would you describe someone’s face when they are expressing skepticism? A noun.”
Develop languages as mentioned above. In one example, I am trying to make sure when I invent a fantasy word for a concept, I am not using any words with french roots. (I am trying to keep the root languages consistent for the fantasy terms. Languages have a certain “sound” that can inform the creation of new languages and anchor readers).
Research things like meteorology, history, etc. For example, what technologies were contemporaneous and what other technologies do they depend on? “How long would it take a fast sailing ship to travel from Ireland to New York in the 1600s at different times of year? What variable affect that?”
Generate names with a particular cultural or linguistic feel. “Give me an old english name for a little boy,” for a single use character or, “Give me ten names that imply smarminess with a Gothic feel” if I’m trying to come up with a name for a more major character.
I have a soft rule that I never upload the actual text of my book for feedback. I keep the actual text of the book out of the LLM’s memory.
I’m not sure where that fits in your model.
Definitely the latter. I would feel stripped of authorship. This isn’t an ethical position, it’s purely emotional / subjective.