The Anatomy of moral agency: A theological and neuroscience inspired model of virtue ethics.

Abstract VirtuosA (‘virtuous algorithm’) is introduced, a model in which artificial intelligence (AI) systems learn ethical behaviour based on a framework adapted from Christian philosopher Dallas Willard and brought together with associated neurobiological structures and broader systems thinking. To make the inquiry concrete, the authors present a simple example scenario that illustrates how a robot might acquire behaviour akin to the virtue of kindness that can be attributed to humans. References to philosophical work by Peter Sloterdijk help contextualise Willard’s virtue ethics framework. The VirtuosA architecture can be implemented using state‐of‐the‐art computing practices and plausibly redescribes several concrete scenarios implemented from the computing literature and exhibits broad coverage relative to other work in ethical AI. Strategies are described for using the model for systems evaluation —particularly the role of ‘embedded evaluation’ within the system—and its broader application as a meta‐ethical device is discussed.