I think we ought to be focusing not on whether it was "real", but on whether it was written by somebody that the user trusts (or maybe there are n trust-hops between reviewer and user). That way users have recourse when they're misled: they can revoke trust in whichever connection exposed them to the misleading review.
Eventually the scammers will be isolated such that they're just paying each other to lie to each other, meanwhile the rest of us can be authentic with each other: we need to learn trust hygiene and bake it into our apps.