How Explainable Medical AI Unlocks Patient Disclosure
(A2025-126303)

AUTHORS

Wansaja BADDEGAMAGE, Université Jean Moulin Lyon 3; Julien Cloarec, iaelyon School of Management, Université Jean Moulin Lyon 3, Magellan; Nour Alrabie, Université Toulouse Jean Jaurès; Maxime Ducret, Hospices Civils de Lyon, Université Claude Bernard Lyon 1, Laboratoire de Biologie Tissulaire et Ingénierie Thérapeutique

ABSTRACT

Explainable artificial intelligence (XAI) is vital for building trust, addressing privacy concerns, and promoting user engagement, particularly in sensitive domains such as healthcare. This study examines the relationships between of XAI, trust, privacy concerns, well-being, and the willingness to disclose personal information. Study 1 reveals that XAI significantly increases trust and decreases privacy concerns, indirectly encouraging information disclosure through trust as a mediator. Study 2 explores the influence of disease severity on user interactions with medical AI, finding that more severe conditions reduce trust, thereby limiting disclosure. Mediation analyses thus underscore trust as the primary mechanism linking XAI to user engagement. These findings underscore the importance of designing transparent, user-focused AI systems that enhance trust and address privacy apprehensions, offering a framework for fostering broader adoption of medical AI technologies.