It seems like a web of trust approach could be useful. For instance, you decide to trust fellow users X to identify who's a creep, and hence automatically adopt their creep block list. Maybe something like this already exists?