In addition it contained info on the recommendations it experienced issued to human reviewers on how to deal with controversial subjects, such as that the AI should "offer to explain some viewpoints of folks and movements", and not present an argument "from its voice" in favor of "inflammatory or harmful" https://lorenzodiotx.getblogs.net/62257064/the-smart-trick-of-chat-gpt-login-that-no-one-is-discussing