Trust in AI Versus Human Clinicians: A Study of Diagnostic Judgements and Transparency
Abstract
Background: Artificial intelligence (AI) is increasingly being considered for use in healthcare, including clinical diagnosis, with potential to support clinicians and address pressures on healthcare systems. However, public trust remains an important consideration in the adoption of AI-based clinical decision-making. This study investigated perceptions of and trust in AI and human clinicians, particularly when their diagnostic judgements were either consistent or conflicting. The study also explored whether concerns regarding transparency and liability influenced preferences for medical AI.
Methods: An experimental vignette and questionnaire-based study was conducted to examine participants' responses to diagnostic judgements made by AI and human clinicians. Vignettes presented scenarios in which AI and human clinicians provided either the same or conflicting diagnostic judgements. Participants' levels of trust and preferences regarding medical AI were assessed, alongside their concerns about transparency and liability.
Results: Participants demonstrated greater trust in human clinicians than in AI clinicians. However, agreement between AI and human clinicians' diagnostic judgements was associated with increased trust in the resulting diagnosis. Participants who expressed greater concern about transparency reported lower levels of trust in medical AI. The findings therefore indicate that perceptions of AI-based diagnosis may be influenced not only by whether the diagnosis is consistent with that of a human clinician, but also by concerns surrounding transparency and accountability.
Conclusions: Trust in AI-based clinical diagnosis appears to remain lower than trust in human clinicians, although agreement between human and AI diagnostic judgements may increase confidence in AI-supported decisions. Concerns regarding transparency may represent an important barrier to trust and acceptance. Developing transparent standards and clearly communicating how AI-based clinical decisions are generated and governed may therefore be important considerations in the adoption of medical AI.
Related
- Clinicians' Trust in AI-Based Clinical Recommendations Across Controlled Primary Care-Style Clinical Vignettes: Nurse-Dominant Experimental Pilot Study.
- Claustrophobia during diagnostic imaging examinations, doctor–patient communication quality, and doctor–patient trust: a cross-sectional mediation and latent profile analysis
- High- versus low-anthropomorphic-cue packages and health-information trust: social presence and source credibility under explicit AI disclosure
- Diagnostic and therapy gaps in BPPV care: a retrospective analysis and feasibility evaluation across a London NHS trust
- CT-Referenced Diagnostic Performance and Complementary Error Patterns of Artificial Intelligence Versus a Single Emergency Physician in Chest Radiograph Interpretation: A Retrospective Paired Diagnostic Accuracy Study
BibTeX
@misc{yang2026trust,
title = {{Trust in AI Versus Human Clinicians: A Study of Diagnostic Judgements and Transparency}},
author = {Ye Yang and Samuel Bennett},
howpublished = {PsyArXiv},
year = {2026},
month = sep,
doi = {10.31234/osf.io/4cbwe_v1},
url = {https://osf.io/preprints/psyarxiv/4cbwe_v1/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Trust in AI Versus Human Clinicians: A Study of Diagnostic Judgements and Transparency
AU - Ye Yang
AU - Samuel Bennett
PY - 2026
DA - 2026/09/21
PB - PsyArXiv
DO - 10.31234/osf.io/4cbwe_v1
UR - https://osf.io/preprints/psyarxiv/4cbwe_v1/
AB - Background: Artificial intelligence (AI) is increasingly being considered for use in healthcare, including clinical diagnosis, with potential to support clinicians and address pressures on healthcare systems. However, public trust remains an important consideration in the adoption of AI-based clinical decision-making. This study investigated perceptions of and trust in AI and human clinicians, particularly when their diagnostic judgements were either consistent or conflicting. The study also explored whether concerns regarding transparency and liability influenced preferences for medical AI. Methods: An experimental vignette and questionnaire-based study was conducted to examine participants' responses to diagnostic judgements made by AI and human clinicians. Vignettes presented scenarios in which AI and human clinicians provided either the same or conflicting diagnostic judgements. Participants' levels of trust and preferences regarding medical AI were assessed, alongside their concerns about transparency and liability. Results: Participants demonstrated greater trust in human clinicians than in AI clinicians. However, agreement between AI and human clinicians' diagnostic judgements was associated with increased trust in the resulting diagnosis. Participants who expressed greater concern about transparency reported lower levels of trust in medical AI. The findings therefore indicate that perceptions of AI-based diagnosis may be influenced not only by whether the diagnosis is consistent with that of a human clinician, but also by concerns surrounding transparency and accountability. Conclusions: Trust in AI-based clinical diagnosis appears to remain lower than trust in human clinicians, although agreement between human and AI diagnostic judgements may increase confidence in AI-supported decisions. Concerns regarding transparency may represent an important barrier to trust and acceptance. Developing transparent standards and clearly communicating how AI-based clinical decisions are generated and governed may therefore be important considerations in the adoption of medical AI.
ER -