Vicarious Sense of Agency in Human-Robot Interaction

Sense of Agency (SoA) is the feeling of having control over one’s actions and outcomes. In humans’ daily life, SoA shapes whether, and how, people feel responsible for their actions, which has profound implications for the organization of human societies. Thus, SoA has received considerable attention in psychology and cognitive neuroscience, which tried to identify the cognitive mechanism underlying the emergence of the individual experience of agency. However, humans are inherently social animals, who are deeply immersed in social contexts with others. Thus, investigations of SoA cannot be limited to understanding the individual experience of agency, as SoA also affects the way people experience others’ actions: this is how SoA becomes “vicarious”. Humans can experience vicarious SoA over another human’s actions and outcomes; however, the mechanisms underlying the emergence of vicarious SoA are still under debate. In this context, focusing on artificial agents may help shed light on the vicarious SoA phenomenon. Specifically, robots are an emerging category of artificial agents, designed to assist humans in a variety of tasks- from elderly care to rescue missions. The present Ph.D. thesis aimed at investigating whether, and under which conditions, robots elicit vicarious SoA in humans in the context of Human-Robot Interaction (HRI). Moreover, we aimed at assessing whether vicarious SoA may serve as an implicit measure of intentionality attribution towards robots. The link between vicarious SoA and intentionality attribution was based on the idea that, in some contexts, humans can perceive robots as intentional agents, and it may “boost” the “vicarious” control that they experience over robot’s actions and outcomes- as well as it happens with other humans. In three studies, we employed the Intentional Binding (IB) paradigm as a reliable measure of implicit SoA. Participants performed an IB task with different types of robots varying in their degree of anthropomorphic features and human-like shape (i.e., the Cozmo robot and the iCub robot). Specifically, our goal was to assess whether the emergence of vicarious SoA in humans was modulated by (1) the possibility to represent robot’s actions using one’s own motor schemes, (2) the attribution of intentionality towards robots, and (3) the human-like shape of the robot. Our results suggested that the interplay of these three factors modulates the emergence of vicarious SoA in HRI. In conclusion, the findings collected in the present thesis contribute to the field of research on the vicarious SoA phenomenon in HRI, providing useful hints to design robots well-tailored to humans’ attitudes and needs.