You do make a very good point about the potential for abuse. My vision for this would be be a dispassionate system -- driven by AI and untouched by politics-- that could create/assign the social score. A model could be created by the science community with inputs from the legal system.