AI Agents and User Acceptance of Recommendations: Effects of Different Visual Representations of AI on User Acceptance
João Diogo Oliveira Araújo · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT) · 2026
As artificial intelligence (AI) systems become integral to digital personalization, understanding the psychological mechanisms that drive user trust and adoption is increasingly critical. This research investigates the impact of different visual representations of AI-powered recommendation agents, non-visual, human-like, and robot-like, on user trust and intention to adopt AI-generated recommendations. Grounded in the literature on personalization, trust, anthropomorphism, animacy, and intrusiveness, a conceptual model was empirically tested through an online experimental survey conducted in Portugal, yielding 197 valid responses. Participants were randomly assigned to one of three visual conditions and assessed on constructs relating to trust and adoption. The findings reveal that trust in competence consistently and significantly predicts adoption intention across all conditions. Perceived animacy serves as a strong antecedent of competence trust, while expected personalization positively influences both competence and integrity, though its effect diminishes under the robot condition. Anthropomorphism and intrusiveness do not significantly affect trust formation. These results indicate that perceptions of expertise and responsiveness outweigh visual human-likeness in shaping user acceptance. The study advances theoretical understanding of trust in AI-mediated personalization and offers practical implications for designing recommendation systems that emphasize competence, accuracy, and transparency rather than anthropomorphic design features.