I doubt the fake ratings companies are using non-employees to generate these ratings. The fake ratings seem to follow many formulaic word patterns and constructions along with certain "buzz" words. I suspect most of these ratings text portions (scripts) are simply generated by computer apps and passed along to employees who are instructed to insert the scripts into the fake reviews they create for the company on behalf of their clients. So we are likely not talking about a company having to reveal confidential information on subscribers who get paid to make fake reviews. I suspect all or most of the fake reviews generated by these companies for their clients are strictly made by the employees who first create many fake user names/IDs then add fake reviews under these IDs using the scripts.