How Personalization and Green AI Shape Empowerment, Trust, and Well- Being in AI Food Recommenders
(A2026-131914)

AUTHORS

Farnazsadat Hosseinchi Sarrafi, Toulouse School of Management; Lars Meyer-Waarden, Toulouse School of Management Research -CNRS University Toulouse 1 Capitole; st�phane salgado, Toulouse School of Management - Universit� Toulouse 1 Capitole

ABSTRACT

Artificial intelligence (AI) recommender systems increasingly guide consumers toward healthier and more sustainable food choices. Yet, personalization in AI-based recommendations demands intensive computation and data processing, raising environmental concerns. This study examines the personalization–Green AI trade-off by exploring how personalization (high vs. low), Green AI framing (Green vs. non-Green), and recommendation orientation (sustainable vs. healthy food) shape psychological empowerment, cognitive trust, well-being, and behavioral intention. A 2 × 2 × 2 between-subjects online experiment (N = 394) showed that both high personalization and Green AI framing independently enhanced empowerment, which in turn increased trust, well-being, and intention to use. Green behavior strengthened responses to Green AI, while privacy concerns showed no significant effect. Results reveal a value-congruence mechanism: Green AI framing fosters trust in personalized systems among eco-conscious users. The study advances understanding of sustainable AI adoption and offers implications for designing responsible, value-aligned recommender systems.