Nothing particularly new, since none of the cases around this have concluded. If I were to guess, I'd say the output of an LLM isn't copyrightable (it's not th

Nothing particularly new, since none of the cases around this have concluded. If I were to guess, I'd say the output of an LLM isn't copyrightable (it's not the creation of a human), unless it's a verbatim copy of some copyrighted training data in which case it belongs to the authors of the work(s) used in training. This creates the most annoying combination of legal problems around using it, so by Murphy's Law it must be correct!