Furthermore, it contained info on the suggestions it experienced issued to human reviewers on how to cope with controversial topics, like which the AI should "present to explain some viewpoints of individuals and actions", and never supply an argument "from its voice" in favor of "inflammatory or unsafe" subjects (although https://adolfh319hqz8.mysticwiki.com/user