What’s in a Name? Implications of AI Roles and Mind Perception for Human-AI Teams
研究了AI角色（队友、支持、工具）如何通过心智感知（能动性和意识体验）影响用户接受度，发现称AI为“队友”能提升心智感知和接受度，但强制标签可能带来负面效果。
Objective We examine how artificial intelligence (AI) roles (teammate, support, tool) shape acceptance via the mind perception subdimensions of agency and conscious experience. Background Adoption of the term AI “teammate” has outpaced evidence of the term’s impact. Such reframing does not guarantee that the benefits of human teamwork will extend to human-AI teams. Understanding the potential benefits and risks of reframing AI as a teammate is essential for guiding effective integration. Methods Across three studies—one survey of employees who use AI at least weekly (Study 1) and two experimental vignette designs (Studies 2-3)—we examined the relationship between AI roles and acceptance using both technology-centered (technology acceptance model, cognitive trust) and human-centered (affective trust) perspectives. This design distinguished role perception from presentation (Study 1 versus Studies 2-3). Results As hypothesized, “teammate” was associated with higher AI mind perception relative to less collaborative roles. While perceiving AI as a teammate showed only positive relationships with acceptance, presenting AI as a teammate was negatively related to some acceptance outcomes after controlling for mind perception. Further, agency and conscious experience were differentially related to technology- and human-centered outcomes, supporting the importance of integrating both perspectives. Conclusion Shifting AI roles from tool to teammate may enhance acceptance via mind perception, but forcing labels may have hidden costs. Applications Collaborative AI roles such as “teammate” should be strategically implemented and aligned with intended mind perception and outcomes. Intermediate labels such as “support” may be more appropriate for describing the roles of today’s AI systems.