Inner Speech for Ethics by Design: From Moral Judgment to Artificial Wisdom

As artificial agents increasingly operate in complex real- world environments, ethical decision-making cannot rely solely on predefined rules or static principles. Situations characterized by am- biguity, cultural variation, and competing values require systems ca- pable of context-sensitive judgment. This paper proposes a cognitive architecture for Ethics by Design grounded in the concept of artificial wisdom, understood as the capacity of autonomous agents to deliber- ate about morally relevant situations and act appropriately across het- erogeneous normative contexts. The approach integrates two comple- mentary foundations: the Aretai model, which operationalizes Aris- totelian phronesis (practical wisdom) into functional capacities, such as moral perception, moral deliberation, emotional regulation, and moral motivation, and the mechanism of inner speech, conceived as a form of functional consciousness that enables self-monitoring and reflective reasoning. Within the proposed architecture, inner speech functions as the internal workspace where ethical considerations are articulated as evaluative statements, allowing the agent to interpret morally salient aspects of a situation, weigh alternative courses of action, regulate affective responses, and maintain continuity between judgment and behavior. This internal discourse also enhances trans- parency and trust in human–robot interaction by making the moti- vations underlying the agent’s decisions intelligible.Finally, the ar- chitecture supports learning and adaptation through reflective feed- back loops, enabling artificial agents to refine their practical judg- ment through experience. By combining insights from virtue ethics, cognitivescience, and robotics, the paper outlines a concrete frame- work for implementing ethically aware artificial systems capable of context-sensitive moral reasoning.