Adoption of artificial intelligence tools among pharmacy students in Syria: patterns of use, educational perceptions, and institutional barriers

BACKGROUND: Artificial intelligence tools are reshaping global higher education, yet their deployment and systemic implications within resource-constrained, conflict-affected settings remain poorly characterized. This study examined artificial intelligence adoption prevalence, usage patterns, perceived educational benefits, self-reported usability, attitudes, social norms, and institutional support structures among undergraduate pharmacy students in Syria. METHODS: A cross-sectional survey was administered between January and February 2026 to 295 pharmacy students across multiple private and public Syrian universities. Data collection utilised a five-construct psychometric instrument. To enhance analytical depth, independent-samples t-tests, one-way Analysis of Variance with post-hoc Tukey's tests, and multiple linear regression modelling were applied to evaluate variations across student subgroups and determine relational dependencies. RESULTS: Overall, 86.8% of participants utilised artificial intelligence tools for academic purposes, with ChatGPT emerging as the dominant platform (96.5%). Core academic use cases included concept explanation (83.2%), drug information retrieval (70.3%), and practice question generation (57.0%). While students reported positive perceived educational benefits (mean = 3.60/5.00) and favourable learning attitudes (mean = 3.66/5.00), institutional support was critically deficient (mean = 1.55/5.00). Formal institutional guidance (93.6%) and training (95.3%) were virtually absent, forcing 95.9% of students to rely entirely on self-directed learning. Inferential analysis revealed significant variations by curricular seniority; fifth-year students demonstrated higher artificial intelligence self-efficacy and usability scores than junior counterparts (P < 0.001). Construct perceptions did not vary significantly by gender, though private university students reported higher institutional support than public university peers (P = 0.036). CONCLUSIONS: Syrian pharmacy students have autonomously integrated artificial intelligence into their academic routines at rates comparable to high-income settings, yet they do so in the complete absence of institutional scaffolding. The co-occurrence of high adoption and positive attitudes alongside deficient critical verification skills exposes a distinct risk profile: students may develop misplaced confidence in automated pharmacological outputs without possessing the evaluative competencies to intercept factual errors. This pattern demands urgent curricular interventions across low- and middle-income countries.