What kind of protections are in place? Could there not be a situation where a fake user confirms his or her identify with the help of a fake agent?