Nursing students’ trust in artificial intelligence (AI) clinical recommendations: A multicenter cross-sectional study of risk-benefit perceptions across Saudi Arabian universities

Abstract

Background Artificial intelligence (AI) is increasingly embedded in clinical decision-making, making trust in AI-generated recommendations essential for safe and effective healthcare delivery. Nursing students, as future practitioners, are critical stakeholders in this transition; however, their trust in AI within rapidly digitizing educational and clinical contexts in Saudi Arabia remains underexplored, despite national digital health initiatives aligned with Vision 2030. Objective This study examined nursing students’ trust in AI-driven clinical recommendations across Saudi universities and investigated how perceived risks, perceived benefits, AI education, and academic progression influence trust development. Methods A multicenter cross-sectional survey was conducted between September and November 2025 across five Saudi nursing institutions (three all-male, one all-female, and one mixed-gender university). A total of 622 undergraduate nursing students completed validated instruments measuring functional and ethical trust, perceived benefits, perceived risks, applicability, and barriers to AI learning. Multiple linear regression and moderation analyses were performed using robust standard errors. Results Although students reported highly positive attitudes toward AI (mean = 3.94/5.00) and strong perceived benefits (mean = 3.98/5.00), overall trust in AI-generated clinical recommendations was low (mean = 2.28/5.00), indicating a trust–optimism paradox. Perceived risks positively predicted trust (β = 0.136, p = 0.010), suggesting that awareness of AI limitations may enhance confidence. This association was substantially stronger among students with prior AI education, whereas senior students demonstrated a weakened risk–trust relationship (β interaction = −0.274, p = 0.009), potentially reflecting skepticism shaped by limited clinical AI exposure. Conclusions Nursing students’ trust in AI is dynamic and influenced by educational exposure and clinical experience. Transparent, structured AI education integrating technical and ethical dimensions is associated with calibrated trust and support for responsible AI adoption in clinical practice. These findings have direct implications for nursing education policy, suggesting that integrating transparent, evidence-based AI curricula into digital health education frameworks can support workforce readiness and responsible technology adoption aligned with national healthcare transformation initiatives.