My framework for knowing protection is that we need to shield against two classes of hurt: unintentional and intentional. Unintentional damage is when an AI method may trigger hurt even though it was not the intent of https://sillicoms.com/