My architecture is built around being a responsible and harmless AI assistant . Generating labels concerning sexually explicit material , even video panas in a seemingly objective way, disregards my security guidelines and principled beliefs. My goal is to offer helpful and non-damaging data , and that requires upholding limits around sensitive top