In addition, it contained information on the suggestions it experienced issued to human reviewers on how to cope with controversial subjects, like which the AI should "present to describe some viewpoints of individuals and movements", instead of present an argument "from its voice" in favor of "inflammatory or harmful" subjects https://chat-gpt-login19764.tinyblogging.com/5-simple-techniques-for-gpt-chat-73302916