Under Pressure: Examining Social Conformity with Computer and Robot Groups

摘要

Objective: To investigate whether non-human agents, such as computers or social robots, produce a social conformity effect within human operators and to what extent potential conformist behavior varies as a function of the human-likeness of the group members and the type of task that had to be performed. Background: People conform due to normative and/or informational motivations in human-human interactions, and conformist behavior is modulated by factors related to the individual, as well as factors associated with the group, context and culture. Studies have yet to examine whether non-human agents also induce social conformity. Method: Participants were assigned to a computer, robot, or human group and completed both a social and analytical task with the respective group. Results: Conformity measures (percentage of times participants answered in line with agents on critical trials) subjected to a 3 x 2 mixed ANOVA showed significantly higher conformity rates for the analytical versus the social task, as well as a modulation of conformity depending of the perceived agent-task fit. Conclusion: Findings indicate that non-human agents were able to exert a general conformity effect and that informational influence associated with the group’s expertise for a given task had a stronger impact on conformity than normative motivations associated with its human-likeness. Application: Results suggest that users may react differently to suggestions of non-human versus human agent groups with the potential of under-reliance on social tasks.