Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

How does copyright work in that case? Can't we reasonably assume that GPT is reciting stories from random web pages, which belong to their author?

AFAIK GPT-3 was not trained only on public domain data?



> Can't we reasonably assume that GPT is reciting stories from random web pages, which belong to their author?

Are there cases of GPT reciting stories? Most content I've seen from it has seemed original enough that I couldn't find a source by searching parts.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: