What's more, it contained info on the suggestions it experienced issued to human reviewers on how to handle controversial subjects, like which the AI must "provide to explain some viewpoints of people and movements", rather than offer an argument "from its voice" in favor of "inflammatory or perilous" topics (even https://chatgptlogin31086.blazingblog.com/29846565/detailed-notes-on-login-chat-gpt