Furthermore, it contained info on the recommendations it experienced issued to human reviewers on how to tackle controversial topics, like which the AI must "present to describe some viewpoints of individuals and actions", instead of offer an argument "from its voice" in favor of "inflammatory or unsafe" subject areas (even https://www.petskyonline.com/aquatic-supplies/