So, RoyalRoad made a post about AI generated content and it gave me ideas.

I wanted to see it for myself if it was any good and the result is:

ChatGPT was able to create three honest versions of my intended prologue chapter after three hours of feeding it my universe, character and scenario, going as far as extrapolating information I had not feed it to fill the blank.

I found the result not good enough but still terrifying so the end result is a 50/50.

But indeed, I can confirm that this tool is dangerous and might do more than simply assisted writers in the near future. But it work the charm against the writer block.

I just gave it my dream unfiltered and unable to turn it into a workable chapter and suddenly, I had something that still needed a lot of work but was actually workable.

For the sake of experiment and transparency I'm going to make a spreadsheet here about the IA and my number count as well as the time it took me to feed the AI + my own rewriting in a neat spreadsheet that would appear right there:

AI words count AI feeding time Human words count Human rewriting time Prologue Chapter 2412/2849 (84, 6%) ~3h +437 ~2h30 Chapter 1 2317/2746 (84,3%) ~6h +429 ~1h30 Chapter 2 2685/2645 (99% *) over 9h N/A * ~30*min

(* Chapter 2: I only removed and replaced words that did not fitted the context. Like tea, which was anachronistic. Or replace sea by lagoon, because the MC literally lived all her life in a port, so she should not be impress at the sight of the sea. Replaced if with lagoon, which might be exotic enough for her. I also add a single sentence of context about Melianna’s whereabouts during the pirate attack and the obvious reasons for the skip over what would otherwise and been must-have content.)

If you come across this story on Amazon, it's taken without permission from the author. Report it.

Bear in mind that it is not a regular novel project. I'm doing it a bit for curiosity, a bit for science and a lot more because I'm bored with no novel to read. My actual novel are 100% my own and I only use IA assistant for grammar checking (grammarly) because I can't afford to pay for it.

If this experiment is a success (and a bad news for writer) chatGPT should grow more and more efficient at writing this novel according to my prompt the more I feed it.

If this experiment is a failure (and a good news for writer) chatGPT should show little to no growth in efficiency or even starting to produce aberrant results that would need a lot of rework on my part.

Finally, I working a full time job so update would be erratic. I'm going to priorities my life over this experiment. So no schedule or anything. Hopefully, there is enough words for the submission so see you in a few hours for this experiment starting event.