Algorithmic Nudging, Transparency and Platform Trust: An Investigation of User Decision Quality In Online Recommendation

Algorithmic nudges have been widely deployed to influence individuals and collective behaviors including unintended results. Drawing on literature of nudging, trust, and cognitive theory, this study focuses on two characteristics of recommendation badges in the world’s largest e-commerce platform and its small-scale counterpart. We conduct two randomized experiments to test two hypotheses:(1) users will have higher decision quality with transparent badge than with non-transparent badge and this effect is more salient when the recommendation preference does not match the user preference,(2) users will have higher decision quality on a platform which they have higher trust compared to one they have less trust in and this effect is more salient when the recommendation preference does not match the user preference. This study will contribute to the literature on nudge, recommendation agent/systems and bring ethical implications to the use of AI in IS research.

In The Proceedings of The 26th Annual SAIS Conference, Gulf Shores, AL, USA, 2024