DETAILED NOTES ON CHAT GPT

What's more, it contained info on the suggestions it had issued to human reviewers regarding how to handle controversial topics, like that the AI should "offer to explain some viewpoints of people and actions", rather than give an argument "from its voice" in favor of "inflammatory or hazardous" subjects (even though it may still "explain arguments