Reasoning with Recommender Systems? Practical Reasoning, Digital Nudging, and Autonomy

Abstract One of the core tasks recommender systems is often defined as follows: Find good items. Recommender are thus designed to support our decision-making by helping us find way around online world and guiding things we want. However, relying on has a profound effect because they structure environment in which make choices. In this contribution, I examine how affect practical reasoning whether pose threat autonomy, i.e., what influence have capacity for making own argue that basic requirement integrating automated recommendations autonomous consists being able identify rationale behind recommendations: only if understand why presented with certain it possible them be integrated into preserves autonomy.