It has been my belief that training a system on content to generate other content, is making a derivative of that content. With most free licenses this is not allowed without mentioning the source (all of them) and in case of normally copyrighted work it's just plainly forbidden.

I hope this case, and the others they mention, succeed.

BBC News - Game of Thrones author sues ChatGPT owner OpenAI
bbc.co.uk/news/technology-6686…