How robots can help: communication strategies that improve social outcomes

摘要

Offering help is a socially precarious venture. A robot that mimics human help-giving communication might end up supporting or might end up offending the help recipient. This thesis begins by observing the varied linguistic strategies human help givers use and their subsequent effects on help recipients. With this understanding, this thesis experimentally observes reactions to robot helpers in comparison to human helpers, looking closely at the influence of help messages on impressions. This experiment provides evidence that imperative statements from a robot are perceived to be controlling, in much the same way as humans using imperative statements. But when particular politeness strategies are used, robots are judged to be even less controlling than people. This thesis improves our understanding of human help-giving communication, offers guidance in the design of sensitive robot helpers, and argues for the continued investigation of advantageous differences between social responses to technology and social responses to people.