>borrowing a book is not creating a COPY of the book. you are not taking the pages, reproducing all of the text on those pages, and then giving that reproduction to your friend.

That's not what's happening with training AI models either though.

check out my other comment in this thread about derivative works.

https://news.ycombinator.com/item?id=42282443

OpenAI are taking copies of people’s data. some of that is copyrighted data.

that’s copyright infringement.

an LLM is a tool to create derivative works from the data OpenAI has copied without permission (when considering only copyrighted works and nothing public domain).

derivative works can also be considered copyright infringement in some cases.

how the tool functions is irrelevant for the most part. how copy right infringement occurs doesn’t matter. only that it does.