Relational vs. group self-construal: Untangling the role of national culture in HRI

As robots (and other technologies) increasingly make decisions on behalf of people, it is important to understand how people from diverse cultures respond to this capability. Thus far, much design of autonomous systems takes a Western view valuing individual preferences and choice. We challenge the assumption that Western values are universally optimal for robots. In this study, we sought to clarify the effects of users' cultural background on human-robot collaboration by investigating their attitudes toward and the extent to which people accepted choices made by a robot or human assistant. A 2×2×2 experiment was conducted with nationality (US vs. Chinese), ingroup strength (weak vs. strong) and human vs. robot assistant as dimensions. US participants reported higher trust of and compliance with the assistants (human and robot) although when the assistant was characterized as a strong ingroup member, Chinese as compared with the US subjects were more comfortable. Chinese also reported a stronger sense of control with both assistants and were more likely to anthropomorphize the robot than were US subjects. This pattern of findings confirms that people from different national cultures may respond differently to robots, but also suggests that predictions from human-human interaction do not hold universally.