Artificial intelligence and online health information use, self-diagnosis, and physician trust among adults in Saudi Arabia: A cross-sectional study

Abstract

Introduction: Artificial intelligence (AI) tools and online health information are increasingly integrated into population health behavior in Saudi Arabia, raising concerns about self-diagnosis, cyberchondria, and trust in physicians.
Objectives: To assess online health information–seeking behavior (OHISB), cyberchondria severity, and physician trust among adults, and to examine their interrelationships.
Methods: In this cross-sectional study, an online questionnaire was distributed nationwide from May to June 2026, covering sociodemographic characteristics, OHISB (5 items, α = 0.83), the Cyberchondria Severity Scale–Short Form (CSS-12; α = 0.89), and physician trust (5 items, α = 0.71). Responses from 1280 adults were analyzed using descriptive statistics, one-way analysis of variance, Pearson correlation, and multiple linear regression.
Results: The mean OHISB score was 3.38 ± 0.70, the mean CSS-12 score was 31.28 ± 7.23 (28.4% high severity), and the mean physician trust score was 3.90 ± 0.51. Sociodemographic variables were not associated with physician trust. Physician trust correlated inversely with OHISB (r = −0.245, P < 0.001) and CSS-12 (r = −0.148, P < 0.001), and both scores were highest in the lowest trust tertile (F = 28.14 and F = 12.97; P < 0.001). In multiple linear regression, OHISB and CSS-12 were independent negative predictors of physician trust (adjusted R² = 0.062).
Conclusion: Online health information seeking, including AI-based tools, and cyberchondria severity were independently associated with lower physician trust, underscoring the need for digital health literacy programs and clear guidance on the responsible use of AI-based health information.