DETRIMENTAL TRUST IN AUTOMATION: HOW CONVERSATIONAL ROBO ADVISORS LEVERAGE TRUST AND MISCALIBRATED RISK TAKING
(A2020-64337)

AUTHORS

Christian Hildebrand, University of St. Gallen; Anouk Bergner, University of St. Gallen

KEYWORDS

Investment Automation; Robo Advisors; Human-Computer Interaction

ABSTRACT

The current work examines how automated financial advisory, in which consumers engage in a dynamic, dialogue-based interaction with a conversational robo advisor, can alter perceptions of trust, the evaluation of the firm, and ultimately consumer financial decision making. Building on recent work on relationship formation between humans and inanimate objects, we provide evidence that conversational robo advisors cause greater levels of affective and cognitive trust compared to non-conversational robo advisors, and a more benevolent evaluation of the financial services firm. We demonstrate that this benevolence attribution can lead consumers to select objectively incorrect portfolio recommendations and accept costly investment products that invoke larger annual management fees. These findings have important implications for research on trust formation between human actors and non-human objects and the emerging research on the potentially detrimental effects of anthropomorphizing machines in the context of consumer financial decision making and consumer protection.