The precedent case in the US formed a legal consensus that "AI" content can't be copyrighted, but it may also contain unlicensed/pirated IP/content.

Thus, one should not contaminate GPL/LGPL licensed source code with such content. The reason it causes problems is the legal submarines may (or may not if they settled out of court with Disney) surface at a later date, as the lawsuits and DMCA strikes hit publishers.

It doesn't mean people won't test this US legal precedent, as most won't necessarily personally suffer if a foundation gets sued out of existence for their best intentions/slop-push. =3