Talk by Johanna Seibt on Robophilosophy and Integrative Social Robotics
Title: The Problem of Artificial “Social Others”: How Much Sociality Do We Need for Hybrid Intelligence?
Abstract: Co-working, co-creating, cooperation, collaboration—social interactions among humans take many different forms even when the single actions involved are functionally similar: What we take ourselves to be doing together, depends on how we do it together. This raises a difficulty for applications of AI and robotics: in the social interaction scenarios we currently envisage for future public use of robots and AI, we need to design them as artificial social agents. But how much sociality is needed? In order to answer this question, I argue, against current dogma in HRI, we need to replace the theoretical construct of “anthropomorphizing”: when we attribute capacities of coordination to artificial agents we sociomorph them, treating them as different types of social agents without attributing them necessarily human capacities. I introduce the descriptive framework OASIS (OASIS: Ontology of Asymmetric Social Interactions) which can be used in to explore in detail which forms of sociomorphing are conducive to an envisaged application. Distinguishing ten levels of sociality and five degrees of simulation, OASIS allows for highly differentiated descriptions of human social interaction with robots and AI’s, which are needed to help ‘users’ adjust their expectations of reliance and to perform ethical risk assessments of AI’s with variable level of social interaction style. In conclusion I sketch the use of the OASIS framework for the purposes of “Integrative Social Robotics”, a value-driven R&D model for the production of culturally sustainable applications.
Professor Seibt is the director of the Aarhus University Robophilosophy Research Unit, which conducts interdisciplinary philosophical and empirical research on human-robot interaction and social robotics.
This talk is part of the lecture series, “Co-Constructing Intelligence,” which focuses on the human-machine interface, and on questions about how humans and machines can learn together and work together to acquire new knowledge and skills. The talk is co-hosted by the CoAI JRC, CITEC, and the Joint Artificial Intelligence Institute (JAII).
The talk will take place in hybrid format in CITEC room 1.204 and via zoom at this zoom link.