Explainability and Preference Manipulation of Algorithmic Nudge @ INFORMS 2023

This paper investigates the impacts of digital nudging on customer purchase decisions. Drawing on literature of nudge and anchoring effect, this study proposes two types of nudges based on the transparency level: ambiguous badge (ex., Amazon’s Choice) and specific badge (ex., Best Seller). We further hypothesize that that specific badge will be less likely to manipulate user’s preferences than ambiguous badge under the condition of preference mismatch. This study will contribute to the nudge literature, recommendation system and dark side in IS and bring ethical implication to the use of AI/ML model and algorithmic nudging. The empirical result can be applied to user interface design for lowering the e-commerce product return rate.

Oct 16, 2023 11:15 AM