When Interact with Generative AI: The Impact of Anthropomorphized Generative AI on Productivity
(A2024-119456)

AUTHORS

Kate Jeonghee Byun, NEOMA Business School; Shijin Yoo, Korea University; Michael Haenlein, ESCP Business School; Daegon Cho, KAIST; Youngchan Hwang, KAIST

ABSTRACT

Generative artificial intelligence (AI) has proven beneficial across various domains, offering customized recommendations and functional as a personal assistant. However, there remains limited understanding of how human-like attributes in generative AI impact individuals' productivity. This paper aims to explore the influence of emotional language and visual signals on user productivity. Across four studies, including a field experiment and three lab experiments, the authors demonstrate that encountering emotional language in generative AI can enhance productivity by saving time and improving performance. This effect is expected to be further strengthened when generative AI is anthropomorphized. The authors also intend to uncover the driving forces behind these effects, identifying trust and engagement as mechanisms. The findings suggest that both pre-training generative AI in emotional language and designing anthropomorphism play pivotal roles in boosting productivity.