Randomized Trials Show AI Interventions Lack Consistent Advantage in Nursing Education

Artificial intelligence-supported educational approaches may benefit selected learning outcomes in nursing students, but current randomized evidence does not demonstrate a consistent advantage over established teaching methods. According to a scoping review evaluating the available literature, the integration of artificial intelligence into nursing curricula produces varied results depending on the tool, the comparator, and the specific educational context.

Instructors and programs have deployed a variety of platforms, including generative artificial intelligence tools, interactive chatbots, intelligent tutoring systems, simulation environments, and artificial intelligence-assisted feedback mechanisms.

To identify and analyze relevant trials, the research team structured their scoping review using the Arksey and O’Malley methodological framework. The investigators conducted comprehensive literature searches across four major electronic databases: PubMed, Scopus, Web of Science, and the Cochrane Central Register of Controlled Trials, known as CENTRAL. The search window spanned from July 20 to August 3, 2026, and targeted peer-reviewed, English-language randomized studies that specifically evaluated artificial intelligence interventions in nursing student populations. The authors noted that the review protocol was not prospectively registered prior to conducting the searches.

Following the literature search and screening process, thirty-one randomized studies published between 2023 and 2026 met all eligibility criteria for inclusion in the review. The authors systematically extracted information regarding study characteristics, experimental interventions, comparators, measured learning outcomes, and reported empirical findings.

The thirty-one randomized trials examined an extensive array of artificial intelligence interventions across nursing curricula. The analyzed modalities included generative artificial intelligence applications, conversational chatbots, intelligent tutoring systems, artificial intelligence-supported clinical simulation, virtual reality environments, and artificial intelligence-assisted assessment and feedback platforms.

Across the included trials, the investigators identified a wide spectrum of measured student endpoints spanning cognitive, clinical, affective, and material dimensions. The documented learning outcomes included knowledge acquisition, academic performance, clinical performance, and clinical reasoning, alongside interpersonal skills such as communication and empathy. Researchers also evaluated higher-order thinking, student self-confidence, satisfaction, attitudes toward artificial intelligence, psychological outcomes, and the quality of educational materials generated or utilized during training.

When evaluating educational efficacy, favorable findings were reported across several outcome domains, but overall findings varied substantially according to the intervention tested, the comparator used, the educational setting, and the specific outcome measured. In multiple studies, conventional educational practices or human-supported teaching approaches yielded comparable or superior outcomes when compared with artificial intelligence interventions. The evidence indicated that artificial intelligence did not provide a reliable or universal benefit over standard instructional methods across the mapped randomized literature.

Methodological quality appraisal conducted with the Cochrane Risk of Bias 2 tool revealed pervasive concerns across the current evidence base. Among the thirty-one randomized studies, twenty-eight were judged to have some concerns regarding risk of bias, three were rated as high risk of bias, and none were rated as low risk of bias overall. Furthermore, longer-term follow-up was uncommon across the literature. The authors concluded that heterogeneity in interventions, comparators, outcome measures, and follow-up periods, along with widespread risk of bias, prevents firm conclusions, emphasizing that more rigorous randomized studies with longer follow-up are needed.

Subject of Research: Randomized evidence on artificial intelligence educational interventions and learning outcomes in nursing students

Article Title: Mapping randomized evidence on artificial intelligence-based educational interventions and learning outcomes in nursing students: a scoping review

Article References: Abdi, N., Salehnasab, C., Dasdar, S., Mirzaee, M. S., & Mehrayin, A. (2026). Mapping randomized evidence on artificial intelligence-based educational interventions and learning outcomes in nursing students: a scoping review. BMC Medical Education. https://doi.org/10.1186/s12909-026-10834-2

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10834-2

Keywords: Nursing education, Artificial intelligence, Scoping review, Randomized controlled trials, Generative AI, Chatbots, Intelligent tutoring systems, AI-supported simulation, Virtual reality, AI-assisted feedback, Learning outcomes, Clinical reasoning, Arksey and O’Malley framework, Cochrane Risk of Bias 2, Academic performance, Self-confidence, Higher-order thinking