The thing is, you have to index the pieces of metadata or otherwise make them discoverable, and then decide which parts you trust.

This would need some kind of signing/trust-distribution scheme, something like namespaces ("YIFY can only approve movie and show releases because they only do movie and show releases").

It would also need way to blacklist malicious metadata (automatic scanners that publish lists of files with viri?).

It's very much non-trivial as far as I can see.

That's ignoring the copyright issue, which can mostly be ignored if you somehow make it impractical to prosecute the distributors of the metadata. But I think (part of) the metadata will still be subject to copyright lawsuits, in the context of "the right to be forgotten" and fair-use safe harbors.