Barriers and facilitators to artificial intelligence adoption among nursing students: a mixed-methods study

Background: Artificial intelligence offers transformative potential for nursing practice, yet significant barriers hinder adoption. While existing research has documented challenges among practicing nurses, limited evidence exists regarding nursing students' perspectives; they are the generation that will shape artificial intelligence's future role in healthcare. Objective: To explore barriers and facilitators to artificial intelligence adoption among nursing students using a mixed-methods approach, examining relationships between technological readiness, ethical concerns, and perceived usefulness of artificial intelligence in nursing practice. Design: Explanatory sequential mixed-methods study combining quantitative surveys with qualitative semi-structured interviews. Setting: College of Nursing, Taibah University, Medina, Saudi Arabia (November 2024-March 2025). Participants: 348 nursing students across academic levels 3-8 participated in the quantitative phase (response rate: 72.5%), with 17 students purposively selected for qualitative interviews. Methods: Validated instruments - the Technology Readiness Index 2.0, an adapted Perceived Usefulness Scale, and an Ethical Concerns Scale - were administered online. Kendall's tau correlation and partial proportional odds modeling identified predictors of perceived usefulness. Qualitative data underwent thematic analysis using Braun and Clarke's framework. Trustworthiness was addressed through investigator triangulation, an audit trail, reflexive memoing, and member-checking. Mixed-methods integration followed a joint-display framework to examine convergence between quantitative and qualitative findings. Results: < 0.001) showed strong positive associations with perceived artificial intelligence utility. Paradoxically, moderate ethical concern predicted higher perceived usefulness (adjusted odds ratio for perceiving low utility = 0.19, 95% CI: 0.09 to 0.40, from the partial proportional odds model). From the qualitative analysis, we revealed universal concern about deskilling (100% of interviewees), data privacy risks (94%), and erosion of human connection in patient care. Participants proposed shared decision-making models where artificial intelligence provides recommendations while nurses retain final clinical authority. The lack of nursing-specific artificial intelligence tools emerged as a critical barrier. Conclusions: Nursing students in this Saudi sample demonstrated nuanced perspectives on artificial intelligence adoption, characterised by cautious optimism alongside critical awareness. The paradoxical relationship between ethical concern and perceived utility challenges traditional technology acceptance models, suggesting deeper engagement fosters appreciation of both opportunities and challenges. We have underscored the importance of tailored educational strategies addressing technical competencies alongside ethical reasoning and professional identity formation. As this generation of digitally fluent students transitions into nursing practice, the perspectives of this sample offer insights for developing artificial intelligence integration approaches that preserve nursing's humanistic core while leveraging technological capabilities.