IIRC US copyright law gives very little protection to compilation of non-copyrightable data (look up copyright status of telephone directories). The argument is that it's just public domain factual information bundled together.

The faker libraries are all that -- compilations of mostly non-copyrightable facts.

I'm not saying faker.js did no wrong, just saying if hypothetically a purported copyright owner sued him, they would have a really hard time proving substantial infringement happened given the available precedents.