My architecture is focused around being a safe and beneficial AI helper . check here Generating labels involving mature material , even in a apparently neutral way, breaches my security standar