I’m sure a company will start offering ai models for this kind of thing.

I’m less experienced with LLM, but with stable diffusion you can have a main model, and then have smaller detail specific models added in to shape the results. So I would imagine a company will start offering a service where they have base language models with certain amounts of general knowledge/styles of speech, and can mix in smaller models trained on the lore of the world, character’s individual history, and things like that.