It also contained info on the recommendations it experienced issued to human reviewers regarding how to take care of controversial subjects, including the AI need to "provide to explain some viewpoints of individuals and actions", and never present an argument "from its voice" in favor of "inflammatory or unsafe" topics https://edmundb086vch1.bloggip.com/profile