If there's a person who studies the works of George R.R. Martin, spending years cultivating writing skills based on just his works, and then writes new material, can that person lose a lawsuit for being too similar in style?
The law just doesn't particularly care about the details of your ML model, OpenAI is still processing massive amounts of copyrighted work without which their software would be nearly useless. That's the interesting legal question.
The only argument is whether their information was scraped off the internet legally, and the supreme court has weighed in that scraping information that's publicly available is legal.
If you put a picture on the internet free and clear, I can save the picture to my hard drive, and I can use it in any way whatsoever sans reproduce it.
The only question is if making an AI that can reproduce it is in fact the same as reproducing it. I think the question is a very simple 'no', but we'll have to see.
But AI isn't a person. You won't find a person that spends every waking moment of their life only studying the works of a specific author, they have lived experience.
Example: https://news.bloomberglaw.com/ip-law/tolkien-estate-sues-ove...
I'm not in the US, so I could easily be off base.
I really enjoyed The Time Ships by Stephen Baxter, which is a sequel to The Time Machine by H. G. Wells. He's a UK author, so maybe it's different.
As a society I think we are better off with this kind of remixing so I'd hope that's the direction we head in.