Understanding users’ AI manipulation intention: An empirical investigation of the antecedents in the context of AI recommendation algorithms

Abstract

This study examines antecedents that drive platform users to manipulate artificial intelligence (AI) recommendation algorithms. Based on the persuasion knowledge model (PKM), survey data collected from YouTube and Instagram users reveal that AI manipulation intentions are positively affected by persuasion knowledge about AI and perceived interactivity. Perceived interactivity is associated with higher perceived benefits and lower perceived costs of AI manipulation, consequently affecting manipulation intentions. A multivariate analysis of variance shows variations in intentions to use different types of AI manipulation behaviors among users with varying levels of persuasion knowledge. The research contributes to the PKM and AI-human interaction literature.