Two authors sued OpenAI, accusing the company of violating copyright law. They say OpenAI used their work to train ChatGPT without their consent.
Say I see a book that sells well. It’s in a language I don’t understand, but I use a thesaurus to replace lots of words with synonyms. I switch some sentences around, and maybe even mix pages from similar books into it. I then go and sell this book (still not knowing what the book actually says).
I would call that copyright infringement. The original book didn’t inspire me, it didn’t teach me anything, and I didn’t add any of my own knowledge into it. I didn’t produce any original work, I simply mixed a bunch of things I don’t understand.
That’s what these language models do.