In addition, it contained info on the recommendations it experienced issued to human reviewers regarding how to handle controversial topics, which includes which the AI should "offer you to describe some viewpoints of individuals and actions", rather than offer an argument "from its voice" in favor of "inflammatory or risky" https://avinnocriminalconvictions68902.blogmazing.com/35130995/the-best-side-of-avin-convictions