The numbers tell a striking story. Only two relevant studies were published in 2019, but by 2025 that figure had climbed to 82, with almost two-thirds of all included studies, 64.6 percent, appearing in just the last two years of the review window. This surge tracks closely with the COVID-19 pandemic, which forced nursing schools worldwide to replace in-person clinical placements and simulation laboratories with digital alternatives almost overnight. What began as an emergency workaround, the review suggests, has hardened into a mainstream pedagogical approach that continues to accelerate even as campuses have fully reopened.

Geographically, the research boom is heavily concentrated. South Korea led the field, producing 21.2 percent of the included studies, followed by the United States at 18.7 percent, with China and Taiwan together contributing 11.1 percent. That East Asian dominance is one of the review’s most notable findings, and it cuts both ways: it reflects serious national investment in simulation-based training in the region, but it also raises questions about how well findings from these settings generalize to nursing programs in other countries, health systems, and cultural contexts. The authors flag this geographic concentration as a key constraint on the evidence base.

Technologically, the field is dominated by immersive experiences. Interactive immersive virtual reality, typically delivered through head-mounted displays that let students manipulate virtual instruments and interact with virtual patients, and 360-degree or 180-degree video-based virtual reality together accounted for 63.1 percent of the included studies. Screen-based virtual simulation, which runs on ordinary computers and requires no special hardware, made up 23.7 percent, while augmented reality, which overlays digital information onto the real world, accounted for a comparatively modest 4.0 percent. The distribution suggests that despite the cost and logistical burden of headsets, educators and researchers have gravitated toward the most immersive formats available, perhaps betting that presence and interactivity translate into deeper learning.

The methodological profile of the literature is equally revealing. Quasi-experimental designs were the most common study type at 27.8 percent, followed closely by randomized controlled trials at 22.7 percent. That nearly a quarter of the field consists of randomized trials is a point in its favor, since such designs are the gold standard for establishing causal effects. Yet the review also uncovered persistent weaknesses: small sample sizes, a scarcity of longitudinal follow-up that would show whether learning gains persist beyond the immediate post-test, and inconsistent outcome measures that make it difficult to compare results across studies or pool them in meta-analyses.

What are nursing educators actually measuring when they deploy these technologies? Three outcomes dominate the literature. Self-confidence and self-efficacy were assessed in 44.4 percent of studies, knowledge acquisition in 42.4 percent, and learner satisfaction in 39.4 percent. The review reports that favorable findings were commonly reported across these outcomes, meaning students generally emerged from virtual simulation feeling more capable, knowing more, and enjoying the experience. But the authors are careful to note an important limitation of their own work: this scoping review did not chart the direction or magnitude of effects, and no formal methodological quality appraisal was undertaken. In other words, the review maps where the field has been, not how strong the ground is.

That distinction matters for anyone hoping to draw practical conclusions. A scoping review, conducted here according to the Joanna Briggs Institute methodology and reported against the PRISMA-ScR checklist, is designed to chart the breadth of a literature rather than weigh the quality of individual studies. The team searched four major databases, PubMed, CINAHL via EBSCOhost, the Web of Science Core Collection, and Scopus, on 15 January 2026, with two reviewers independently screening records and extracting data and a third adjudicating disagreements. The protocol was retrospectively registered on the Open Science Framework, and the authors published their complete data extraction matrix, covering all 198 studies, as supplementary material, an unusually transparent move that allows other researchers to scrutinize and reuse the dataset.

The technical breadth captured in the review also hints at where the field is heading. Beyond the headline categories of virtual, augmented, and mixed reality, the included studies span a wide range of nursing education domains, from fundamental clinical procedures to complex scenario-based decision making. The authors standardized heterogeneous labels from the original papers into consistent categories for technology type, delivery format, outcome measures, sample-size bands, and educational domain, a data harmonization effort documented in their supplementary files. This kind of infrastructure matters because a literature growing this fast risks fragmenting into incompatible silos of terminology and measurement unless someone takes the time to build a common map.

The review’s conclusions are measured but pointed. Virtual simulation has been implemented across a broad range of nursing education contexts, and the mapped studies predominantly reported positive learning outcomes. However, the authors identify four constraints that warrant attention: limited longitudinal follow-up, geographic concentration in East Asia, small samples, and inconsistent outcome measures. Their prescription is standardized, multi-site research, ideally with agreed-upon instruments and designs that follow students long enough to show whether virtual simulation produces durable clinical competence rather than a temporary boost in test scores or confidence.

For nursing schools weighing investment in headsets, simulation software, and faculty training, the review offers both encouragement and caution. The sheer volume of recent research, and the consistency of positive findings on confidence, knowledge, and satisfaction, suggests that virtual simulation is no passing fad but a durable fixture of clinical education, one that proved its worth when the pandemic severed students from real wards. At the same time, the field’s reliance on small, short-term, regionally concentrated studies means that the boldest claims about virtual simulation, that it can replace or even surpass traditional clinical experience, still outrun the evidence. The next phase of this research story, the authors imply, will be less about proving that the technology can engage students and more about demonstrating, with rigorous and comparable methods, that it makes them better nurses.

Subject of Research: Use of virtual simulation technologies in nursing education

Article Title: Virtual simulation in nursing education: a scoping review

Article References: Dong, H., Zhou, C., Xing, L., & Li, D. (2026). Virtual simulation in nursing education: a scoping review. BMC Nursing, 25(1), Article 780. https://doi.org/10.1186/s12912-026-05295-3

Image Credits: AI Generated

DOI: 10.1186/s12912-026-05295-3

Keywords: virtual simulation, nursing education, virtual reality, augmented reality, scoping review, nursing students, educational technology, simulation-based training, COVID-19, clinical education, self-efficacy, knowledge acquisition

The study, led by Yan Peng and Mengqi Li of Fuyang Medical School, responds to a transformation that is reshaping drug production worldwide. Modern pharmaceutical manufacturing is moving away from traditional batch-based production toward digitalized, continuous, data-driven systems driven by Industry 4.0, smart manufacturing, and regulatory modernization. Technologies such as Quality by Design (QbD), Process Analytical Technology (PAT), Manufacturing Execution Systems (MES), electronic batch records (EBR), digital twins, and artificial intelligence-assisted process optimization are becoming essential components of production and quality assurance. Yet conventional pharmaceutical education remains largely theory-centered and textbook-based, leaving graduates with limited exposure to real production scenarios, digitalized manufacturing systems, and data-driven decision-making.

The mismatch is not trivial. Practical teaching in pharmaceutical production is constrained by high costs, safety requirements, and the regulatory complexity of GMP-compliant facilities, so students often graduate with insufficient practical competency and digital manufacturing literacy. To address this, the research team designed a curriculum module that transformed scattered intelligent manufacturing topics into a coherent learning system spanning six content areas: QbD and PAT for real-time quality monitoring; MES and EBR for digital production scheduling and data traceability; digital twins and virtual simulation; AI-assisted process optimization; GMP compliance, validation, and data integrity; and enterprise-based innovation projects built on authentic industrial problems.

The pedagogical engine of the module is a four-integration teaching model. Curriculum content was integrated with industrial standards, aligning teaching with GMP requirements and enterprise production practices. Classroom learning was integrated with production scenarios, using real enterprise cases, deviation records, and quality risk situations. Physical practice was integrated with virtual simulation, compensating for the constraints of pharmaceutical production training with safe, repeatable digital environments. Finally, knowledge learning was integrated with innovation projects, requiring students to apply theory to team-based tasks such as electronic batch record optimization, PAT-based process monitoring, and digital twin modeling. The whole framework was supported by an industry-education co-cultivation mechanism in which enterprise experts helped design the curriculum, delivered case-based lectures, mentored projects, and co-evaluated student outputs alongside university instructors.

To test whether this elaborate framework actually worked, the team ran a quasi-experimental pretest-posttest control group study with 92 undergraduate students in pharmaceutical engineering and related majors during the 2025 academic term. Two intact classes of 46 students each were assigned to the intervention or control group. Because random assignment at the individual level was not feasible, the researchers rigorously tested baseline equivalence across all outcome measures and used analysis of covariance (ANCOVA) with pretest scores as covariates to control for residual baseline differences. Both classes were taught by the same instructor to minimize instructor effects. The intervention group received the eight-week module with two contact hours per week plus project activities; the control group received conventional lectures and textbook-based instruction covering similar basic topics but without systematic modules on MES, digital twins, AI optimization, or enterprise mentoring.

The measurement instruments were engineered with unusual care for an education study. The 30-item knowledge test covered QbD, PAT, MES, EBR, digital twins, AI-assisted optimization, GMP, validation, and data integrity, achieving a Scale-Level Content Validity Index of 0.92 and Cronbach’s alpha of 0.82. Practical competency was assessed with a blinded two-rater performance rubric (inter-rater reliability ICC = 0.86), while digital manufacturing competency was measured through scenario-based tasks such as interpreting MES data showing a trend of increasing tablet hardness across three batches and proposing data-driven optimization strategies. Learning engagement was measured with a 12-item scale covering behavioral, cognitive, and emotional engagement (alpha = 0.90). Parallel pretest and post-test versions drawn from the same item bank correlated at r = 0.85 to minimize practice effects.

The results were dramatic. In knowledge acquisition, the intervention group achieved a mean learning gain of 22.8 points versus 13.8 points in the control group, a large effect (Cohen’s d = 0.95). Practical competency gains were 24.5 versus 15.2 points (d = 0.98). Learning engagement rose more in the intervention group (1.13 versus 0.61 points, d = 0.76). But the standout result was digital manufacturing competency: the intervention group gained 25.1 points against only 11.7 points for the control group, a very large effect of d = 1.25, with the intervention group’s advantage reaching 13.4 points, the widest gap of any outcome. ANCOVA confirmed that group membership remained a significant predictor of post-test scores for all four outcomes after controlling for baseline, with the strongest effect on digital manufacturing competency (F = 32.56, p < 0.001, partial eta squared = 0.27).

Qualitative evidence reinforced the numbers. Students in the intervention group completed team-based innovation projects on authentic problems including electronic batch record optimization, PAT-based monitoring, digital twin modeling, and data integrity risk assessment, earning an average project score of 86.3 out of 100. Enterprise mentors rated the projects with the same rubric, and the correlation between mentor and instructor scores was r = 0.89, indicating strong convergence. Six mentors rated the module highly on relevance to workplace needs (4.3 out of 5) and student readiness for industrial practice (4.0 out of 5). One mentor noted that a student project on electronic batch record optimization was directly applicable to the company’s current digital transformation initiatives. Student satisfaction with the virtual simulation and digital twin tools was high (4.15 out of 5 for perceived utility), with overall module satisfaction at 4.08.

Thematic analysis of student feedback identified four major themes: enhanced understanding of real industrial pharmaceutical manufacturing, improved digital manufacturing awareness, increased learning motivation, and a need for more training in data analysis. One student reflected that previously they only understood pharmaceutical manufacturing from textbooks, but through the virtual production line and enterprise cases they realized how process parameters, quality attributes, equipment, and data systems are connected. Another observed that future pharmaceutical manufacturing requires not only knowledge of pharmaceutics but also the ability to understand data, digital systems, and intelligent equipment. The recurring complaint was that the AI optimization component was useful but difficult, particularly for students with limited programming or statistics backgrounds, suggesting future iterations should include more scaffolded training in data analytics and computational thinking.

The study’s limitations are acknowledged by the authors: the quasi-experimental design lacked individual randomization, the sample came from a single institution with two intact classes, the eight-week intervention did not assess long-term retention, and some assessment tools were developed specifically for this study. Still, the framework offers a transferable template for institutions worldwide grappling with the same challenge, since pharmaceutical manufacturers everywhere are adopting digitalized production systems, electronic quality management, and data integrity requirements. The four-integration teaching model and industry-education co-cultivation mechanism may be adaptable to other applied disciplines facing digital transformation, from biotechnology to chemical engineering. For a field where graduates must integrate pharmaceutical science, engineering practice, digital technologies, and regulatory thinking, the message is clear: the curriculum of the future cannot be built from textbooks alone.

Subject of Research: Education

Article Title: Development and evaluation of an intelligent pharmaceutical manufacturing curriculum module under industry–education integration: a quasi-experimental study

Article References: Peng, Y., & Li, M. (2026). Development and evaluation of an intelligent pharmaceutical manufacturing curriculum module under industry–education integration: a quasi-experimental study. Discover Education, 5(1), Article 1179. https://doi.org/10.1007/s44217-026-02148-8

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02148-8

Keywords: pharmaceutical manufacturing, digital twins, curriculum innovation, industry-education integration, virtual simulation, competency-based education, GMP, quasi-experimental study, Development, evaluation, intelligent, pharmaceutical

Fiber post-core restoration is a particularly instructive test case for educational technology. The procedure is not a single maneuver but a chain of interlocking decisions and actions: the clinician must assess whether the remaining tooth structure can support a post, select the appropriate post system, determine the correct length and diameter of the post space, handle the root canal environment without compromising the seal, and then build and shape the core so that the eventual crown has a sound foundation. Each of these steps depends on case-specific judgment, which means a student who has memorized the sequence can still fail if they cannot reason through an individual patient’s presentation. Traditional preclinical teaching, which relies heavily on bench-top simulation and limited instructor contact time, struggles to give every student enough supervised practice in both dimensions at once.

The platform described in the study was designed as a closed-loop system rather than a simple skill trainer. According to the authors, most existing virtual simulation systems in dental education function primarily as procedural rehearsal tools, allowing students to practice hand movements and instrument sequences on a simulator without any structured support for the decision-making that surrounds those actions. The new platform integrates three components intended to address that deficiency. The first is a dynamic AI virtual patient consultation, in which students interact with a simulated patient whose presentation drives the diagnostic and treatment-planning phase of the exercise. The second is a domain-specific procedural question-and-answer assistant that provides targeted guidance during the operative phase, effectively standing in for the one-on-one instruction that is scarce in crowded preclinical courses. The third is a standardized, rubric-based medical record evaluation workflow, which requires students to document their clinical decisions and then assesses that documentation against explicit criteria.

The study enrolled two independent cohorts of fourth-year dental students, thirty-eight in 2024 and forty-two in 2025, in a pre-post design. Before training, students completed a clinical case analysis assessment intended to measure clinical reasoning. After the full training cycle, they repeated the case analysis assessment and also completed a simulator operation assessment measuring procedural performance at course completion. The researchers then compared pre- and post-training scores within each cohort and examined the relationships between reasoning scores, procedural scores, and the change in reasoning performance over the course.

The headline quantitative result was a statistically significant improvement in clinical case analysis scores in both cohorts. In the 2024 group, median scores rose from 69 before training to 81.5 afterward; in the 2025 group, the median climbed from 71 to 83. Both increases were highly significant, with p values below 0.001. On its face, that pattern suggests that students emerged from the AI-supported curriculum substantially better at working through the diagnostic and planning decisions that precede a fiber post-core restoration. The consistency of the improvement across two separate years of students adds a degree of reassurance that the observation was not a one-off artifact of a single unusual cohort.

Yet the authors are notably careful about how far that conclusion can be pushed, and their caution is a useful lesson in reading educational research. The study found a strong negative correlation between pre-training scores and score gains: students who started low improved the most, with a correlation coefficient of -0.879 in 2024 and -0.789 in 2025. The researchers point out that this pattern should be interpreted with caution because it may reflect statistical artifacts rather than genuine educational effects. Regression to the mean, the tendency of extreme baseline values to move toward the average on retesting, naturally produces the largest apparent gains among the weakest performers. Ceiling effects compress the possible improvement for high scorers. And there is a mathematical coupling problem inherent in correlating a baseline score with a change score that is computed from that same baseline. In other words, the striking correlation says less about the platform than about the arithmetic of pre-post designs.

A second set of correlations was more straightforwardly encouraging. Post-training clinical reasoning scores and post-training procedural operation scores were strongly and positively associated in both cohorts: r = 0.927 in 2024, r = 0.722 in 2025, and r = 0.840 across the combined sample, all statistically significant. This suggests that, at the end of the course, students who reasoned well through clinical cases also tended to perform well on the simulator. The authors frame this as a positive association between decision-making and procedural performance rather than proof of causation, but the alignment matters for curriculum design. It indicates that the platform’s dual emphasis on reasoning and technique did not leave the two strands of competence developing independently; students strong in one tended to be strong in the other.

The subjective side of the evaluation was similarly positive. Students completed a study-specific questionnaire covering learning interest and satisfaction, self-perceived outcomes, career expectations, and acceptance of AI in their training. The questionnaire’s face and content validity were assessed through internal review by the three study authors who developed it, an important caveat the authors themselves flag, since this constitutes an internal-team assessment rather than independent external expert validation. Internal-consistency reliability was assessed for the multi-item dimensions. Across the two cohorts, students reported high course satisfaction with a mean rating of 4.26 on a five-point scale. Their endorsement of AI’s utility was even stronger, at a mean of 4.29, with the platform’s perceived ability to compensate for limited one-on-one teacher guidance standing out as a particular strength. Career expectations were the highest-rated construct of all, at a mean of 4.34, suggesting that students saw the training as relevant to their professional futures rather than as an academic exercise.

Those perception findings speak to a real structural problem in dental education. Preclinical courses typically pair large student cohorts with a limited number of instructors, which means each student receives only fragments of individualized feedback during hands-on sessions. An AI assistant that is available at every step of a simulated procedure, answering procedural questions on demand, effectively scales the guidance that a single instructor can provide. The students’ strong ratings of that function suggest they experienced the technology as filling a genuine gap rather than as a gimmick. At the same time, satisfaction surveys measure how students feel about a course, not whether the course made them better clinicians, and the authors are explicit that the questionnaire was built for this specific investigation rather than validated externally.

Indeed, the most distinctive feature of the paper may be the discipline of its own conclusions. Because the study used an uncontrolled pre-post design, with no comparison group receiving conventional instruction, and because procedural performance was assessed only after training with no pre-training baseline, the authors state that the findings should be interpreted as preliminary implementation-related observations rather than evidence of the platform’s effectiveness, educational equity, or improvement in procedural skills. That framing is refreshingly honest in a field where educational technology is often promoted with far stronger claims than the data can bear. What the study does establish is that the platform can be implemented across successive cohorts, that clinical reasoning scores rose after the curriculum, that reasoning and procedural performance moved together, and that students embraced the experience. The work was supported by a virtual simulation teaching reform project at Nankai University, and the authors declare no competing interests. Whether the closed-loop approach, with its virtual patient consultations, procedural AI assistant, and rubric-based record evaluation, ultimately outperforms traditional preclinical teaching will require the controlled comparative trials that this implementation study deliberately stops short of claiming. For now, it offers a detailed and cautiously reported blueprint for how AI might be woven into the earliest stages of surgical training without overstating what the evidence shows.

Subject of Research: AI-enhanced virtual simulation for preclinical dental education in fiber post-core restoration

Article Title: A closed-loop AI virtual simulation platform for preclinical fiber post-core restoration: two-year prospective evaluation of learning outcomes and student perceptions

Article References: Li, Y., Chen, C., & Liao, X. (2026). A closed-loop AI virtual simulation platform for preclinical fiber post-core restoration: two-year prospective evaluation of learning outcomes and student perceptions. BMC Medical Education. https://doi.org/10.1186/s12909-026-10510-5

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10510-5

Keywords: artificial intelligence, virtual simulation, dental education, prosthodontics, fiber post-core restoration, clinical reasoning, preclinical training, medical education, student perceptions, BMC Medical Education, Nankai University, competency-based education

The study, led by researchers affiliated with the Second Xiangya Hospital of Central South University, was conducted between August and December 2025. Teachers at 16 medical colleges across central, eastern, and western China voluntarily forwarded a standardized anonymous survey link to nursing students. Eligible participants were full-time associate degree, undergraduate, or master’s students who had used an adaptive virtual simulation platform at least twice in their current-semester courses. After applying strict quality controls—including a minimum completion time of 120 seconds, patterned-response detection, and an embedded attention-check item—2,016 valid questionnaires remained from 2,030 returned.

The platform under study, developed by Beijing Oubeier Software Technology Development Co. Ltd., included more than 50 nursing-skill modules covering fundamentals of nursing and medical-surgical nursing, with each simulation task typically lasting 20 to 30 minutes. Importantly, the researchers describe its adaptive functions as rule-based rather than genuinely intelligent: the system compared students’ recorded operational steps, decision pathways, and response times against predefined task rules and performance thresholds to trigger branch-specific feedback and adjust scenario difficulty. No individual platform logs were linked to the survey responses.

The sample was predominantly female (83.48 percent) and aged 19 to 24 (88.39 percent), comprising 1,580 undergraduate students, 380 associate degree students, and 56 postgraduates. The researchers measured four core constructs using validated instruments: the System Usability Scale (SUS), the extraneous cognitive load subscale of Leppink’s Cognitive Load Scale, an adapted nine-item version of the Flow in Education Scale (EduFlow-2), and the nine-item Utrecht Work Engagement Scale for Students (UWES-S). All scales showed acceptable to excellent reliability, with Cronbach’s alpha values ranging from 0.749 to 0.899.

The descriptive picture was sobering for platform designers. The mean usability score translated to roughly 56.4 on the conventional 0-to-100 SUS scale—below the widely used benchmark of 68 that signals above-average usability. In other words, students on average perceived the system as only moderately usable, even as they reported moderate levels of load, flow, and engagement.

Structural equation modeling revealed a clear pattern of statistical associations. Greater perceived usability was associated with lower extraneous cognitive load (standardized coefficient β = −0.497), the load imposed by confusing interfaces and unclear instructions rather than by the learning material itself. Higher extraneous load was in turn associated with lower flow experience (β = −0.481)—the absorbing state in which learners concentrate fully under balanced challenge—and greater flow was strongly associated with higher learning engagement (β = 0.492). Usability also showed a direct positive association with engagement (β = 0.213).

Bootstrap mediation analysis with 5,000 resamples supported a sequential indirect association: usability → lower extraneous load → greater flow → higher engagement, with a standardized indirect estimate of 0.116 and a 95 percent confidence interval excluding zero. Notably, neither of the two simpler two-step pathways—usability through load alone, or usability through flow alone—reached significance, suggesting that load and flow operate as a linked sequence rather than as independent routes. Extraneous load showed no direct association with engagement once flow was accounted for, implying that its influence on engagement is realized through learners’ experiential state.

The study also examined whether students’ digital readiness—an exploratory index combining prior AI tool use and self-perceived digital competence—changed these relationships. Multi-group analysis, backed by a continuous-variable interaction test, showed that the negative association between extraneous load and flow was stronger among students with lower digital readiness (β = −0.519) than among those with higher readiness (β = −0.374). For less digitally prepared students, a clumsy interface appears to be a more serious barrier to achieving immersive, engaged learning. The authors suggest practical remedies: pre-training orientation sessions, step-by-step guided tutorials, simplified navigation, minimal operational steps, and graduated technical support during initial system use.

The researchers are careful about the limits of their design. All variables were measured concurrently, so the findings describe cross-sectional statistical associations rather than causal effects; temporal ordering and causation cannot be established from this data. Common method bias was formally tested with Harman’s single-factor test and an unmeasured latent method construct approach, both of which suggested it was not a serious concern, but the data remain self-reported. The study also lacked objective learning outcome measures such as Objective Structured Clinical Examination scores, and the purposive dissemination through 16 colleges means the sample cannot be considered nationally representative of Chinese nursing students.

Even so, the implications for nursing education are concrete. As AI-based simulation spreads—publications on AI in nursing grew from 38 in 2014 to 563 in 2024 by one count—institutions should prioritize usability testing and iterative design improvement when procuring educational technology, treat clear navigation and concise instructions as core pedagogical features rather than cosmetic details, and build differentiated support for students with varying digital readiness. The study’s message is that engagement with AI simulation is not simply a matter of supplying advanced technology; it depends on whether the interface lets students spend their limited working memory on nursing knowledge instead of on wrestling with the system itself.

Subject of Research: Cross-sectional associations among AI system usability, extraneous cognitive load, flow experience, and learning engagement in nursing simulation education.

Article Title: AI System Usability and Learning Engagement in Nursing Simulation: Cross‐Sectional Statistical Indirect Associations Through Extraneous Cognitive Load and Flow Experience

Article References: Liu, M., Zhang, P., Wu, Y., Pan, L., & Li, L. (2026). AI System Usability and Learning Engagement in Nursing Simulation: Cross‐Sectional Statistical Indirect Associations Through Extraneous Cognitive Load and Flow Experience. Nursing Open, 13(9), Article e70855. https://doi.org/10.1002/nop2.70855

Image Credits: AI Generated

DOI: 10.1002/nop2.70855

Keywords: artificial intelligence, virtual simulation, nursing education, system usability, cognitive load, flow experience, learning engagement, digital readiness, structural equation modeling, cross-sectional survey, nursing students, educational technology

The research team, led by Xuebing Jing of Zibo Central Hospital and Yanyan Men of Qilu Medical University’s College of Nursing, used a statistical technique called latent profile analysis to sort students into subgroups based on their scores on an artificial intelligence literacy scale. Rather than treating AI literacy as a single continuous dimension on which everyone differs only by degree, latent profile analysis asks whether the population contains qualitatively distinct clusters, people whose patterns of knowledge, skills, and attitudes resemble one another more than they resemble the rest of the sample. This person-centered approach has become increasingly popular in educational research because it recognizes that competence is multidimensional and that different combinations of strengths and weaknesses may require different remedies.

The analysis converged on a three-profile solution that fit the data better than any alternative. The smallest group, accounting for 24.57 percent of participants, showed low AI literacy across the measured dimensions. A moderate-literacy group made up 37.22 percent of the sample, while the largest cluster, 38.21 percent of students, demonstrated high AI literacy. That so many students cluster at the high end might seem reassuring, but the researchers emphasize the heterogeneity: nearly one in four future nurses enters clinical training with a weak grasp of AI concepts, even as AI tools are being embedded in electronic health records, triage systems, and patient monitoring platforms worldwide. A curriculum pitched at the average student will inevitably leave the low-literacy quarter behind while boring the high-literacy group.

What determines which profile a student lands in? The researchers ran multivariate logistic regression to identify factors significantly associated with membership in each group, and six predictors emerged. Internship experience mattered, suggesting that exposure to real clinical environments, where AI tools increasingly appear in daily workflows, builds practical AI understanding that classroom instruction alone does not. Formal AI-related training was another significant predictor, an intuitive but important confirmation that structured coursework moves the needle. The frequency of daily AI use in students’ personal lives also predicted higher literacy, hinting that familiarity with chatbots, recommendation systems, and smart devices outside school translates into institutional confidence. Participation in innovation and entrepreneurship activities, such as research projects or design competitions, similarly distinguished the higher-literacy profiles.

Perhaps the most intriguing predictors, however, were psychological rather than experiential. Critical thinking, measured with the California Critical Thinking Disposition Inventory, and ethical sensitivity, assessed with the Ethical Sensitivity Questionnaire for Nursing Students, both significantly predicted profile membership. This finding reframes AI literacy as more than a technical skill set. Students who habitually question assumptions, weigh evidence, and reason through moral dilemmas are the same students who engage more deeply with AI, perhaps because they recognize both its power and its limits. For educators, the implication is that ethics and critical thinking courses are not adjuncts to AI training but integral components of it, shaping the cognitive dispositions that allow technical knowledge to take root.

The quantitative survey only tells half the story. To understand what students actually want from AI education, the team conducted 15 semi-structured interviews and analyzed the transcripts using conventional content analysis, a qualitative method in which codes and themes emerge from the data rather than being imposed in advance. The interviews revealed that educational needs spanned two broad domains: the format of teaching and the content of learning. On the format side, students expressed a clear preference for interactive, case-based learning over passive lectures. They wanted to work through realistic clinical scenarios in which AI tools play a role, discussing in small groups where an algorithm’s suggestion should be followed, when it should be questioned, and how to explain decisions to patients and families.

The students also identified specific technologies they believed should be woven into the curriculum. Virtual simulation, which allows learners to practice clinical decision-making in safe, repeatable digital environments, was viewed as a natural fit for AI education. Generative AI tools, the same large language model systems that have transformed public discourse since their emergence, were seen as useful teaching instruments in their own right, capable of supporting practice and exploration, provided their limitations are made explicit. This appetite for hands-on engagement with the very technologies reshaping healthcare suggests that students are not afraid of AI; they want structured, supervised opportunities to master it before they encounter it unsupervised on a hospital ward.

On the content side, students’ requests were concrete and pragmatic. They wanted foundational instruction in how AI actually works, the basic principles underlying machine learning and algorithmic prediction, so that they could understand what these systems can and cannot do rather than treating them as opaque oracles. They wanted clinical application skills: how AI is used in nursing practice today, from early-warning scores to documentation support, and how to operate and interpret such tools competently. And they wanted sustained coverage of legal and ethical guidelines, including data privacy, accountability, and the boundaries of algorithmic decision-making in patient care. Notably, students in the high-literacy profile emphasized one further need that the researchers found particularly telling: support for professional identity formation, the process of understanding what it means to be a nurse in an era when machines participate in diagnosis and monitoring.

The study’s conclusions point toward a differentiated approach to AI education in nursing. Because literacy is heterogeneous and shaped by multiple, identifiable factors, the authors argue that educational strategies should be tailored to specific profiles. Students in the low-literacy group may need foundational, confidence-building instruction with abundant scaffolding, while high-literacy students may be better served by advanced applications, innovation opportunities, and reflective work on professional identity. Across all groups, the integration of practical clinical experience, deliberate critical thinking and ethics training, and diverse teaching formats, from case-based discussion to virtual simulation, appears essential. As AI continues its advance into every corner of healthcare, this research offers a rare, evidence-based map of where the nursing workforce stands, and a practical blueprint for bringing every student, not just the technologically fluent majority, up to the demands of intelligent medicine.

Subject of Research: AI literacy profiles and educational needs among undergraduate nursing students

Article Title: Profiles of AI literacy and educational needs of nursing students: a mixed‑method study

Article References: Jing, X., Men, Y., Wang, Q., Wang, J., Li, R., & Zhou, X. (2026). Profiles of AI literacy and educational needs of nursing students: a mixed‑method study. BMC Medical Education. https://doi.org/10.1186/s12909-026-10381-w

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10381-w

Keywords: artificial intelligence, AI literacy, nursing education, nursing students, latent profile analysis, mixed methods, educational needs, critical thinking, ethical sensitivity, virtual simulation, case-based learning, healthcare