Beyond Strong Social Robotics: A Relational Perspective

Preprint of a paper forthcoming in conference proceedings. This is the author version of a contribution that will appear in:Proceedings of the International Conference on Hybrid Human–Artificial Intelligence (HAR 2025),Lecture Notes in Computer Science (LNCS), Springer, Cham. The final published version may differ from this preprint. --- Abstract This epistemological contribution offers a theoretical reframing of the strong AI debate within the field of social robotics, where the classical strong/weak AI dichotomy has evolved into a tension between two identifiable approaches: strong and weak social robotics. The former seeks to replicate human-like social competencies in artificial agents, grounded in an individualistic model of embodiment; the latter designs robots that evoke social presence without reproducing the deep affective and cognitive processes characteristic of living systems. Despite their differences, both perspectives share the assumption that social competence must be an intrinsic property of the robot. In this article, we challenge this view. Our core proposal is a relational and embodied epistemological framework that conceives sociality not as rooted in the internal (whether “natural” or “artificial”) physiology of individual agents, but as emerging from dynamic processes of human–robot coordination. This theoretical reconceptualization, we argue, also calls for a shift in the ethical domain, which we articulate through synthetic ethics: a context-sensitive, practice-based ethical approach. Moving beyond merely negative ethical framings of social robots, synthetic ethics treats human–robot interaction as a site of ongoing experimentation, where ethical and relational significance are co-constructed. We suggest that this framework enables ethics to engage with the novelty introduced by emerging social machines, and may orient their diffusion in ways that foster human self-understanding, social development, and sustained moral growth.