27 March 2023 in PsyArXiv
Robots are increasingly taking on roles in contexts in which ethical decision making is necessary. In this paper, we offer a set of tools to study one central capacity of future moral robots: to have norm competence and, in particular, to make acceptable decisions when moral norms conflict with one another. Our efforts are in line with growing trends in the community to create and share systematic and replicable methods as well as validated testbeds, measures, and open datasets. Specifically, we propose four principal requirements for systematic research on human responses to robots’ choices to resolve norm conflicts and on their sequelae---human moral judgments, potential loss of trust, and robot mitigations to regain that trust. We also share a series of validated testbeds in the form of moral dilemmas. We detail our procedures and results of a human-subjects experiment (N=1,573) to validate these moral dilemmas as demonstrating high levels of norm conflict and thus well suited to study human-robot interaction in light of such norm conflicts. From these efforts we provide for the community an integrated framework of methodological tools for Moral HRI.
11 March 2024 in OSF Preprints
Building and maintaining trust is critical for continued human-robot teaching and the prospect of robot learning social skills from natural environments. Whereas previous work often explored strategies to reduce system errors, mitigate trust loss, or using enhanced interactivity to induce learning success, few studies have investigated the possible benefits of fully engaged, interactive teaching on human trust. Motivated by the discrepancy discovered from a pair of previous investigations, the studies presented in this paper for the first time directly tested the causal impact of interactivity on the loss and recovery of trust in a human-robot social skills training context. Using a novel paradigm, we experimentally manipulated the mode of interaction that participants were able to engage in (robot supervisor or active teacher) and measured the teachers' changing trust in a 15-trial robot training session, centering on the critical social skills of norm-appropriate behavior. Our results demonstrate that interactive teachers were more resilient to initial trust loss, showed increased reliance from baseline to post-training, and attributed more of the robot's improvement to themselves than did supervisors, even when their robots were imperfectly slower learners.