Artificial intelligence is increasingly used alongside clinicians in medical decision-making, raising questions about public trust. This project examines trust in medical decisions reached jointly by a clinician and an AI system, and how this trust is shaped by the AI’s explanation. Participants will complete an online vignette-based survey presenting a diagnosis reached collaboratively by a clinician and AI, in which the AI’s explanation discloses a limitation in its competence (ability), its capacity to prioritise the patient’s individual needs (benevolence), its fairness and institutional approval (integrity), or discloses none. Drawing on Mayer, Davis and Schoorman’s (1995) framework of ability, benevolence and integrity as the three mechanisms of trust, the primary outcome is trust in the human-AI recommendation, with the primary hypothesis that these conditions differentially affect trust. Data will be collected across countries varying in individualism, uncertainty avoidance, and power distance, to explore whether these trust mechanisms operate similarly across cultures.

Team:

Anna Louise Todsen Supervisor
Çağla Çolak Recruitment and ethics lead
Fabio Leonardi Data and analysis lead
Gabriela de Brito Maciel Theory lead and open science officer
Le Minh Trang Vo Measures and method
Lucija Milović Communication officer and materials lead