INFORMATION TRANSPARENCY AND AI RECOMMENDER SYSTEM: A DOUBLE-EDGED SWORD
(A2025-124440)

AUTHORS

Sara-Maude Poirier, Université du Québec à Montréal; Sylvain Sénécal, HEC Montreal; Pierre-Majorique Léger, HEC Montréal

ABSTRACT

With new global regulations on online privacy, service providers must enhance transparency about how they use consumers’ personal data for AI-driven recommendations. This research aims to understand consumers’ intentions toward transparent AI system depending on whether it uses their personal data for higher personalized recommendations. Our results revealed that transparency alone before the reception of the recommendation can backfire if the AI system is not customizable, i.e., when consumers have no control over the use of their personal data. Conversely, if the AI system does not need the consumer’s personal data to generate personalized recommendations, the consumer’s control over the system is not necessary and transparency only is efficient. This research highlights the need for policymakers to create regulations that operationalize transparent and controllable AI system, alongside its theoretical contributions to the signaling theory, and managerial implications.