You seem to think these models haven't already been trained on pirated versions of this content, for some reason.
That’s not even considering AI crawlers or all the copyright text on archive.org
You seem to think these models haven't already been trained on pirated versions of this content, for some reason.