It also contained information on the recommendations it had issued to human reviewers on how to take care of controversial topics, like which the AI should "offer to explain some viewpoints of people and actions", instead of deliver an argument "from its voice" in favor of "inflammatory or risky" matters https://harryq754scj2.slypage.com/profile