Recent developments, both in the cognitive sciences and in world events, bring special emphasis to the study of morality. The cognitive sciences, spanning neurology, psychology, and computational intelligence, offer substantial advances in understanding the origins and purposes of morality. Meanwhile, world events urge the timely synthesis of these insights with traditional accounts that can be easily assimilated and practically employed to augment moral judgment, both to solve current problems and to direct future action. The object of the following paper is to present such a synthesis in the form of a model of moral cognition, the ACTWith model of conscience. The purpose of the model is twofold. One, the ACTWith model is intended to shed light on personal moral dispositions, and to provide a tool for actual human moral agents in the refinement of their moral lives. As such, it relies on the power of personal introspection, bolstered by the careful study of moral exemplars available to all persons in all cultures in the form of literary or religious figures, if not in the form of contemporary peers and especially leadership. Two, the ACTWith model is intended as a minimum architecture for fully functional artificial morality. As such, it is essentially amodal, implementation non-specific and is developed in the form of an information processing control system. There are given as few hard points in this system as necessary for moral function, and these are themselves taken from review of actual human cognitive processes, thereby intentionally capturing as closely as possible what is expected of moral action and reaction by human beings. Only in satisfying these untutored intuitions should an artificial agent ever be properly regarded as moral, at least in the general population of existing moral agents. Thus, the ACTWith model is intended as a guide both for individual moral development and for the development of artificial moral agents as future technology permits.
展开▼
