Robots in Need? Humans to the Rescue: Unpacking Prosociality toward Machines

Abstract

Die Dissertation ist gesperrt bis zum 20. März 2027 !

The focus of human-robot interaction research has mostly been centered on topics such as robot perception and successful cooperation between humans and robots, with an end objective of implementing robots in day-to-day life in a wide range of fields, such as factories, eldercare, and the medical domain. However, with more and more robots entering common spaces, the variety of interactions is sure to grow. One of these novel interactions refers to humans showing prosociality behaviors toward stranded robots (Hankewitz, 2021; Hinton, 2022; Price, 2022a). As such, a new research endeavor opens up in the form of investigating human to robot prosociality, more specifically, researching under which conditions are people willing/ not willing to be prosocial toward robots. Thus, the objective of this dissertation lies in expanding on human-robot interaction (HRI) literature by investigating varied types of prosociality (willingness to help, helping behavior, and resource sharing perception) through the lens of established HRI variables as well as new ones. It addresses both variables that promote and inhibit prosociality. A total of six experiments with N = 1847 participants were conducted, representing the empirical foundation of this dissertation. Five out of the six experiments were conducted online. Across all experiments, participants showed a form of prosociality. Experiments 1a, 1b, and 1c addressed willingness to help (intention to help), with positive attitudes and context predicting willingness to help rates across all three experiments. The influence of anthropomorphism is non-significant contrary to expectations. Additionally, knowledge about robots further increases willingness to help (1c). Experiment 2 moves from intention to behavior, thus investigating actual helping behavior as influenced by eye-contact (anthropomorphic cue) with a robot. Results show that participants were physically faster in helping the robot when reporting eye contact. Finally, experiments 3a and 3b, address the perception of sharing resources with robots. Adopting a third-person perspective to investigate this matter, results point to the relevance of resource type. That is, sharing of electricity could be seen as more appropriate compared to sharing money with a robot. Attitude factors (robot trustworthiness and risk perception) further influence the sharing perception. The overall results point to high rates of human to robot prosociality. This prosociality is further modulated by a number of variables, the most consistent of which is the relevance of context and framing, as it can both promote or hinder prosocial behavior.