心理学
矛盾心理
社会心理学
应用心理学
认知心理学
沟通
感知
定性研究
情感(语言学)
发展心理学
作者
Yixiu Yu,Ofir Turel,Fred D. Davis,Eric A. Walden
标识
DOI:10.1080/08874417.2026.2656898
摘要
Anthropomorphism is a defining feature of AI voice assistants, with humanlike voice, interactivity, and empathy cues shaping user perceptions. While prior research has examined the benefits and drawbacks of anthropomorphism, its relationship with ambivalence, i.e. the coexistence of positive and negative evaluations, remains underexplored. Drawing on Attitudinal Ambivalence Theory, this study investigates how anthropomorphic cues influence ambivalence and how ambivalence, in turn, shapes attitude and continuance intention. Survey data from users of AI voice assistants (e.g. Alexa, Siri, Google Assistant) show that voice and interactivity cues directly reduce ambivalence, whereas empathy exerts an indirect effect through these pathways. Ambivalence also mediates the effects of these cues on attitude. Moreover, ambivalence exhibits a dual role: it directly reduces attitude while simultaneously strengthening the relationship between attitude and continuance intention. These findings advance IS research by disentangling anthropomorphic design into cue-level mechanisms and by reconceptualizing ambivalence as a context-dependent moderator that both constrains and amplifies user responses. Practically, the study offers design insights for reducing evaluative conflict and fostering more coherent human—AI interactions.
科研通智能强力驱动
Strongly Powered by AbleSci AI