Introduction
Nursing education is facing increasingly complex challenges in competency development. Population aging, chronic disease management, acute and critical care, community-based continuity of care, and digital health services are reshaping the competencies expected of nurses. In clinical practice, nurses are now required not only to master disease knowledge and nursing skills, but also to address patient safety, health information systems, interprofessional communication, and ethical decision-making. As electronic health records, telehealth, smart monitoring devices, and clinical decision support systems are gradually integrated into nursing practice, nursing education must incorporate digital health competencies, information evaluation skills, and technological proficiency into the professional preparation process[
1,
2]. Traditional teaching methods—centered on classroom lectures, textbook study, skill demonstrations, and periodic exams—remain valuable for conveying foundational knowledge and standardized skill training. However, they are relatively inadequate in addressing individual student differences, complex case reasoning, and dynamic clinical scenarios. Therefore, the integration of artificial intelligence (AI) into nursing education is not merely a pursuit of novel teaching formats, but a practical necessity for nursing education to adapt to the digital transformation of clinical practice. Nursing education should cultivate nurses capable of maintaining professional judgment in a technological environment, rather than students who merely know how to use a specific software or platform. The educational value of AI lies in its ability to shift or automate certain repetitive, standardized, and traceable learning support tasks, allowing instructors to devote more energy to training in clinical reasoning, guiding humanistic care, and shaping professional values[
3].
From the perspective of nursing education practice, AI should not be simplistically viewed as a tool that “replaces teachers” or “completes learning tasks on behalf of students.” A more appropriate characterization is that AI can provide an interactive, feedback-driven, and scalable learning support environment for nursing education. Existing research indicates that AI can be applied to various teaching components, including case analysis, virtual simulation, intelligent simulated patients, clinical communication training, automated feedback, and formative assessment, thereby helping to enhance engagement in case-based learning, the timeliness of feedback, and the objectivity of clinical competency evaluation[
4,
5]. If these functions are used only sporadically, AI is likely to remain confined to assisting with lesson preparation or generating teaching materials; only by integrating it into course objectives, practical instruction, and evaluation systems can it better facilitate a shift in nursing education from one-way knowledge transfer to a training model that emphasizes situational learning, evidence verification, and continuous feedback[
5].
For conceptual clarity, this review distinguishes three major categories of AI relevant to nursing education[
6,
7]. The first category is generative AI and conversational agents, including large language model-based tools and chatbots that can generate case materials, answer questions, support reflective writing, and assist students in organizing nursing care plans[
6,
7]. The second category is machine learning and learning analytics systems, which use educational and performance data to support formative assessment, identify learning gaps, and provide early warning for students who may require additional support[
6,
7]. The third category is virtual simulation and intelligent simulated patient systems, which combine scenario-based learning, virtual environments, and adaptive patient responses to support clinical reasoning, communication training, and practical skill development[
8–
10]. These categories are interrelated but should not be treated as identical, because their educational functions, evidence bases, and ethical risks differ[
6,
7,
11]. The major applications, educational value, risks, and safeguards of these AI categories in nursing education are summarized in Table 1.
Accordingly, the purpose of this review is not to advocate the unrestricted adoption of AI in nursing education, but to clarify how AI can be used within pedagogically sound, ethically governed, and professionally supervised educational systems. Specifically, this review aims to: (1) categorize major AI technologies used in nursing education; (2) synthesize their educational value and practical applications in learning support, case-based teaching, simulation training, and assessment; (3) critically examine the strengths and limitations of current evidence; and (4) propose ethical boundaries and curriculum-level safeguards for responsible implementation.
Methods
Review design
This article was designed as a narrative review informed by a structured literature search. A narrative review approach was selected because the purpose of this article was to synthesize conceptual, pedagogical, empirical, and ethical evidence regarding AI in nursing education, rather than to estimate pooled intervention effects. Therefore, this review did not perform a meta-analysis and was not intended to meet the full reporting requirements of a systematic or scoping review. Nevertheless, a structured search strategy was used to improve transparency and reduce arbitrary selection of the literature.
Search strategy
A structured literature search was conducted in PubMed/MEDLINE, CINAHL, Web of Science, and Scopus. The search covered publications from January 2020 to June 2026, with an emphasis on studies published within the last three years due to the rapid development of generative AI and learning analytics in health professions education. Search terms included combinations of the following keywords: “AI,” “generative AI,” “ChatGPT,” “large language model,” “machine learning,” “learning analytics,” “virtual simulation,” “intelligent simulated patient,” “nursing education,” “nursing students,” “clinical reasoning,” “simulation training,” “formative assessment,” “AI literacy,” “academic integrity,” “ethics,” “privacy,” and “patient safety.” Reference lists of relevant reviews and highly cited articles were also screened to identify additional studies closely related to the review topic.
Eligibility criteria
Publications were included if they met the following criteria: (1) written in English; (2) focused on nursing education or health professions education with direct relevance to nursing education; (3) examined AI-related applications in teaching, learning support, case-based learning, simulation training, assessment, AI literacy, academic integrity, or ethical governance; and (4) were empirical studies, randomized or quasi-experimental studies, mixed-methods studies, systematic reviews, scoping reviews, umbrella reviews, ethical guidelines, or high-quality discussion papers. Publications were excluded if they focused only on clinical AI applications without educational implications, described technical algorithm development without a nursing education context, were conference abstracts without full text, were duplicate publications, or had limited relevance to the aims of this review.
Data extraction and synthesis
The included literature was examined according to the following dimensions: AI category, educational application, study design, target learners, major outcomes, reported benefits, implementation barriers, ethical risks, and recommended safeguards. Evidence was synthesized narratively across major educational domains, including personalized learning support, case-based teaching, simulation-based education, formative assessment, AI literacy, and ethical governance. Rather than simply aggregating positive findings, the synthesis compared the maturity of evidence across different AI applications and identified areas in which evidence remains preliminary or inconsistent.
Critical appraisal approach
Because this was a narrative review, no formal quality scoring tool was applied. However, the strength of evidence was appraised narratively by considering study design, sample size, duration of intervention, type of outcome measures, degree of clinical transfer, and relevance to nursing education. Greater interpretive weight was assigned to systematic reviews, scoping reviews, randomized controlled studies, mixed-methods studies with clear educational outcomes, and empirical studies directly involving nursing students or nursing educators. Opinion papers and discussion articles were used mainly to support ethical, conceptual, and governance-related analysis.
Educational value, practical applications, and critical synthesis of AI in nursing education
Educational value of AI in nursing education
Personalized learning support and active learning
The direct value of AI in nursing education is primarily reflected in improving access to learning support. During their studies, nursing students often need to grasp a large amount of specialized knowledge in a short period of time and connect nursing procedures, disease characteristics, and patient scenarios. When faced with key nursing considerations for similar conditions, risk assessments in complex cases, or questions not fully understood outside the classroom, generative AI and intelligent question-and-answer systems can provide students with conceptual explanations, process overviews, and learning prompts. Compared to traditional after-class Q&A sessions, these tools are not constrained by fixed times or locations and can respond to student questions more quickly, thereby alleviating to some extent the learning dilemmas of “being afraid to ask, not knowing how to ask clearly, and waiting for feedback”[
12].
The more significant role of AI is not merely to provide answers but to encourage students to continue probing deeper into the issues. Based on a single case, students can further analyze changes in the patient’s condition, key points of nursing assessment, potential risks, health education, and post-discharge care arrangements. This learning process helps students shift from memorizing facts to understanding problems, comparing solutions, and identifying evidence. Existing research shows that AI-supported case analysis can improve nursing students’ case management performance and increase their engagement in case-based learning[
4]. For students with relatively weaker foundational knowledge, AI can provide necessary conceptual explanations and step-by-step guidance; for students with stronger learning abilities, it can guide them in further developing clinical reasoning skills through comprehensive case studies and reflective questions. From this perspective, AI is better suited as a learning scaffold rather than a tool to replace students’ own thinking[
13].
Although AI-supported learning tools may improve access to explanations and feedback, their educational value should be interpreted cautiously[
6,
7]. Most current studies examine short-term learning outcomes, student satisfaction, or perceived usefulness rather than long-term competence development[
6]. Therefore, generative AI should be positioned as a learning scaffold that encourages questioning, comparison, and evidence verification, rather than as an authoritative knowledge source[
11,
14]. In nursing education, personalized learning support is meaningful only when students are required to justify their reasoning, verify AI-generated responses against reliable sources, and reflect on how patient context affects nursing decisions.
Development of teaching resources and support for course design
AI also offers new support for nursing instructors in course preparation and instructional organization. Nursing education often requires a large volume of case materials, classroom discussion questions, quiz items, and assignment feedback. In teaching scenarios with large class sizes or heavy practical coursework, it is difficult for instructors to consistently track each student’s learning progress within a limited timeframe. AI can be used to generate preliminary lesson plan frameworks, case materials, and formative assessments. It can also assist in organizing student performance data, helping instructors identify common errors and areas of weakness more quickly, thereby improving the efficiency of instructional preparation and feedback[
15].
The value of AI-assisted teaching-resource development should be judged by curriculum alignment rather than by productivity alone[
11]. AI may help instructors draft preliminary cases, quizzes, rubrics, and feedback templates, but these outputs require professional review before classroom use[
11]. Instructors should examine whether AI-generated materials match course objectives, reflect realistic clinical priorities, and avoid misleading or oversimplified nursing recommendations. Therefore, AI may reduce repetitive preparation work, but it does not reduce the educator’s responsibility for content validity, pedagogical coherence, and professional value guidance[
16].
The application of AI in nursing case studies and situational teaching
Generative AI-based case construction and clinical scenario presentation
The core of nursing practice education lies in guiding students to apply theoretical knowledge to specific patient scenarios. Traditional case-based teaching relies on instructors’ long-term experience and manual compilation, which often limits the number of cases, their level of difficulty, and the frequency of updates. Generative AI can rapidly generate draft case studies based on course objectives and students’ foundational knowledge, and supplement contextual information regarding disease progression, nursing assessments, and risk identification. Its value lies not in simply increasing the number of cases, but in helping instructors construct clinical problems at different levels, enabling students to repeatedly compare, evaluate, and refine their approaches within similar scenarios, thereby enhancing the flexibility and reproducibility of case-based teaching[
17,
18].
Case generation must be grounded in authenticity. Nursing cases should not merely list the correspondence between disease names and nursing interventions; they should also present information on the patient’s functional status, psychological reactions, care resources, and treatment preferences. Real-world clinical settings often involve incomplete data, rapidly changing conditions, and uncertain patient choices. If cases are overly standardized, students may tend to reduce nursing judgment to fixed steps. Generative AI can provide case frameworks, but medical facts, nursing processes, and ethical scenarios still require faculty review. When involving real patient data, the information should be rewritten on the basis of de-identification to avoid compromising evidence-based practice and privacy protection requirements due to case generation[
16].
Case analysis and feedback for developing nursing judgment skills
The focus of using AI in clinical case studies should not be limited to “generating a case,” but rather on training students in the nursing decision-making process. Students need to identify primary nursing issues from limited information, prioritize abnormal findings, explain the relationship between nursing goals and interventions, and anticipate risks following the implementation of interventions. By posing a series of follow-up questions, AI can prompt students to explain the rationale behind their judgments, making it particularly well-suited for identifying issues such as leaps in reasoning, insufficient evidence, and unclear nursing priorities. Compared to rote memorization of knowledge, this interactive approach more closely resembles clinical reasoning training and is more effective in cultivating students’ judgment skills when facing complex situations[
4,
12].
Existing randomized controlled studies have shown that AI-supported case analysis can improve nursing students’ case management performance and serve as a supplement to traditional case discussions[
4]. Other studies have applied generative AI to patient simulations, suggesting its potential value in clinical competency training and situational learning[
6]. Accordingly, nursing education can integrate AI into the processes of case preparation, classroom discussion, and post-class reflection. Before class, students use AI to organize case clues; during class, instructors guide students in comparing different nursing care plans; after class, students are required to identify shortcomings in the AI’s responses and revise them by referencing textbooks, guidelines, or classroom content. Through such instructional arrangements, AI ceases to be a tool that completes assignments for students and instead becomes a learning medium used to expose cognitive blind spots, reinforce evidence-based verification, and train nursing judgment.
The evidence for AI-supported case analysis in nursing education is encouraging but still limited[
4,
6]. Randomized and quasi-experimental studies suggest that AI-supported case analysis may improve students’ case management performance, clinical competency, and learning satisfaction. However, most studies have used short intervention periods and measured immediate educational outcomes rather than long-term clinical performance[
4,
6]. In addition, outcomes such as satisfaction, perceived competence, and engagement may not fully reflect students’ ability to transfer reasoning skills to real patient care[
8]. Future studies should therefore include longer follow-up periods, objective clinical performance assessments, and comparisons across different nursing specialties and educational contexts[
6].
AI-assisted simulation-based education and the development of clinical reasoning skills
The application of virtual simulation and intelligent simulated patients in practical education
Nursing education is distinctly practice-oriented, but opportunities for real-world clinical training are not always sufficient or evenly distributed. Some critical care scenarios arise suddenly and involve high risks, which students may not fully experience during their clinical rotations. If education relies entirely on clinical settings, students’ understanding of disease progression, emergency response coordination, and risk communication may remain limited to observation. Virtual simulation and intelligent simulated patients provide a relatively safe training environment for practical education, allowing students to repeatedly practice assessment, judgment, procedures, and communication before entering real clinical settings. Systematic reviews and meta-analyses indicate that virtual simulation plays a positive role in enhancing nursing students’ clinical reasoning, practical skills, and communication abilities, serving as a valuable supplement to traditional simulation training and clinical placements[
9,
19].
Compared to traditional simulation training, AI-supported scenario presentation more closely mirrors clinical processes. Intelligent simulated patients can provide differentiated responses based on students’ questions, interventions, and communication styles, while the case system can gradually release clues about the patient’s condition, enabling students to form judgments based on incomplete information. Students are not faced with predetermined answers but must continuously refine their assessments and intervention plans in response to evolving clinical conditions. Existing research combining AI with virtual reality technology for interprofessional communication training and patient simulation suggests that this approach helps improve student engagement, clinical competence, and situational response capabilities[
20]. Therefore, the educational value of virtual simulation and intelligent simulated patients lies not in replacing real clinical settings, but in providing more stable and controllable training conditions for competency preparation prior to clinical practice.
Compared with generative AI-based case discussion, virtual simulation has a relatively stronger evidence base in nursing education. Systematic reviews and meta-analyses have reported beneficial effects of virtual simulation on clinical reasoning, competence, and communication-related outcomes. However, AI-driven simulated patients and generative AI-based patient simulation remain at an earlier stage of development. Existing studies suggest potential benefits for perceived clinical competence, cultural awareness, communication training, and learner engagement, but the evidence is still insufficient to determine whether these tools produce durable improvements in real clinical performance. Important unresolved issues include scenario authenticity, standardization of evaluation indicators, cultural adaptability of virtual patients, and the degree to which simulation-based learning transfers to bedside practice[
8–
10].
Reflective learning and clinical reasoning training supported by intelligent feedback
Clinical reasoning skills are not developed through a single training session but are gradually honed through action, feedback, and refinement. AI can record students’ operational pathways, decision-making sequences, and reaction times during simulation training, and use this data to highlight issues such as omissions in assessment, inadequate risk identification, or unclear intervention logic. Compared to simply assigning a grade, formative feedback better helps students understand the causes of errors and encourages them to revisit case materials and nursing evidence to revise their judgments. Recent studies on AI-assisted case analysis and clinical competency assessment have shown that such tools can improve case management performance and enhance the timeliness and structure of the evaluation process.
The interpretation of AI-generated feedback requires caution[
5,
11]. Learning analytics systems can record observable behaviors, such as response time, decision sequence, and omission of assessment steps, but they cannot fully interpret the contextual meaning of students’ decisions[
5]. Nursing reasoning is influenced by patient emotion, family dynamics, cultural background, communication tone, and ethical conflict[
21]. These dimensions are difficult to capture through automated indicators alone. Therefore, AI-generated feedback should be used as preliminary process evidence rather than as a final judgment of student competence[
5]. A more appropriate model is to combine AI-supported documentation with faculty-led debriefing, student reflection, and clinical preceptor feedback[
5].
AI-supported teaching workflows and assessment in nursing education
Continuous support throughout the entire teaching process
The impact of AI on the nursing teaching process should not be simplistically interpreted as merely delegating pre-class, in-class, and post-class tasks to a platform. Its more practical value lies in providing teachers with continuous instructional support, thereby creating a tighter connection between course preparation, classroom interaction, and learning feedback. Before class, teachers can use AI to develop course frameworks, design pre-class questions, and prepare supplementary case materials. During classroom instruction, AI can be used to support probing questions, case simulations, and real-time feedback. After class, the system can preliminarily identify common issues in student assignments and quizzes, helping instructors determine which topics require further explanation and which students need additional support[
15].
While this support helps reduce the time spent on repetitive material organization and low-level feedback, it does not mean that teaching can be taken over by a technological platform. Nursing curricula not only impart knowledge and skills but also bear the responsibility of fostering patient safety awareness, professional accountability, and professional values. While AI-assisted approaches can be appropriately introduced for standardized knowledge, procedural content, and simulated cases, content related to nursing ethics, clinical communication, and professional attitudes still requires instructors to provide explanations grounded in real-world contexts. Instructors should always retain control over setting instructional objectives, selecting content, and guiding values. AI can enhance the efficiency of instructional organization, but cannot replace instructors’ grasp of the essence of the nursing profession[
22].
Developing formative assessment and competency-based assessment
AI can also provide richer process-based evidence for nursing education assessment. Traditional assessments primarily rely on theoretical exams, skills tests, and clinical evaluations, which reflect students’ knowledge acquisition or performance outcomes at a specific point in time but struggle to capture changes in their thinking throughout the learning process. Learning analytics and intelligent feedback systems can track students’ performance in case studies, simulation training, and assignment corrections, enabling instructors to identify gaps in students’ knowledge comprehension, clinical reasoning, and clinical readiness at an earlier stage. Related studies indicate that formative assessment can enhance nursing students’ clinical knowledge, skill performance, and self-efficacy, suggesting that continuous feedback supports competency development more effectively than a single summative assessment[
23].
The focus of nursing education assessment is shifting from “whether learning tasks have been completed” to “whether students possess the comprehensive competencies required for clinical practice.” In addition to knowledge and skills, assessment should also address clinical reasoning, patient safety awareness, communication and collaboration, ethical judgment, digital health literacy, and AI literacy. AI can provide data recording, feedback prompts, and preliminary analysis, but it should not serve as the primary basis for final evaluation. Particularly when assessing professional attitudes, ethical sensitivity, and the ability to demonstrate humanistic care, teacher observation, clinical preceptor feedback, and student reflection remain irreplaceable. A more reasonable approach is to use process data provided by AI as one piece of evaluative evidence, combined with teacher evaluations, clinical performance, and student self-reflection to form a competency-based evaluation system[
5,
24].
Despite these advantages, AI-supported assessment remains methodologically and ethically challenging. Learning analytics may help identify learning gaps, but prediction models can be affected by incomplete data, biased training datasets, opaque algorithms, and inconsistent definitions of clinical competence[
11]. Process data should therefore not be used as the sole basis for high-stakes decisions such as course failure, clinical placement readiness, or professional competence judgment. In competency-based nursing education, AI-generated assessment evidence should be triangulated with teacher observation, clinical preceptor evaluation, objective structured clinical examination performance, student reflection, and patient-safety-related indicators.
AI literacy and critical thinking among nursing students and educators
Nursing students’ evidence-based thinking, critical thinking, and AI literacy
With the integration of AI into nursing education, nursing students need to master not only how to operate these tools but, more importantly, how to assess the reliability of their outputs[
25,
26]. Generative AI may contain factual errors, logical inconsistencies, fictional sources, and context mismatches. If students directly use its responses as assignment content or clinical recommendations, it will not only undermine their ability to think independently but may also lead to erroneous nursing judgments. Therefore, when using AI, nursing students should proactively verify its outputs against textbooks, course content, clinical guidelines, and faculty feedback, treating AI as a learning aid rather than an authoritative source of answers[
27].
AI literacy should be integrated into the development of professional nursing competencies. Nursing students need to understand the limits of AI’s applicability, recognizing that while it can help explain concepts, organize thoughts, and prompt questions, it cannot replace evidence-based research or clinical judgment. Students should also learn to formulate questions within a clinical context, incorporating the patient’s basic condition, primary symptoms, nursing goals, and known risks into the questioning process, rather than relying on vague prompts to obtain superficial answers. Existing research suggests that AI tools can enhance students’ learning efficiency and writing performance, but they may also lead to a lack of individualization in care plans and even mask students’ insufficient understanding of the issues at hand[
27]. Therefore, nursing education should neither simply prohibit students from using AI nor allow its unregulated use; instead, the process of using AI should be transformed into training in evidence verification, critical thinking, and professional judgment[
21–
23].
Current studies on AI literacy in nursing education mainly focus on students’ attitudes, readiness, self-efficacy, writing support, and perceived usefulness[
27]. Less is known about whether AI literacy training can produce sustained improvement in evidence verification, clinical reasoning, patient safety awareness, and ethical decision-making[
14]. In addition, students differ substantially in digital access, language proficiency, prior AI experience, and confidence in using AI tools. Nursing curricula should therefore treat AI literacy as a longitudinal competency rather than a one-time technical skill. Future research should evaluate whether structured AI literacy education improves students’ ability to question AI outputs, identify hallucinated information, cite reliable evidence, and maintain independent clinical judgment[
16].
Nursing educators’ technical understanding, content review, and educational responsibilities
The application of AI also places new demands on nursing educators. Educators must not only be well-versed in nursing expertise and educational principles but also understand the fundamental characteristics of AI tools, be able to determine which teaching components are suitable for their use, and identify which content must be explained, demonstrated, and evaluated by the educator in person. AI-generated case studies, nursing recommendations, and learning feedback may appear comprehensive, but they may contain factual inaccuracies, a lack of value judgments, or missing patient context. If instructors lack the necessary discernment, they may inadvertently introduce unverified content into the classroom, thereby affecting students’ understanding of nursing issues[
17,
24].
In teaching practice, instructors should incorporate guidelines for AI use into course requirements. Clear instructions must specify whether AI is permitted in assignments, the extent of its use, whether the process must be disclosed, and whether generated content must be accompanied by verification evidence. For case analyses, nursing plans, and reflective assignments, instructors should also focus on whether students genuinely understand the issues, rather than merely submitting well-written, AI-generated text. AI can help instructors improve the efficiency of lesson preparation and provide students with more practice opportunities; however, instructors must continuously monitor its boundaries of use, ethical risks, and academic integrity requirements. Only under conditions of clear rules, verifiable content, and traceable accountability can AI serve as an effective support for nursing education, rather than posing a hidden risk of undermining students’ professional judgment and academic integrity[
28].
Contextual differences between developed and developing countries
The significance of AI in nursing education differs across developed and developing contexts[
7]. In developed countries, nursing programs are more likely to have access to mature digital infrastructure, electronic health record systems, simulation centers, learning management platforms, AI-supported educational tools, and institutional data-governance mechanisms. These conditions make it more feasible to integrate AI into competency-based assessment, interprofessional simulation, personalized learning, and learning analytics. In such settings, the central challenge is not only whether AI can be used, but how it can be aligned with curriculum standards, accreditation requirements, faculty development, and ethical governance.
In developing countries, AI may have different but equally important educational implications. It may help expand access to teaching resources, support remote learning, generate low-cost case materials, and provide preliminary feedback in settings where faculty resources and clinical placement opportunities are limited. However, these potential benefits are constrained by uneven internet access, limited digital infrastructure, insufficient faculty training, language bias in AI-generated content, lack of locally relevant nursing cases, and underdeveloped data protection systems. AI integration in nursing education should therefore not follow a uniform model. Instead, implementation should be adapted to local infrastructure, regulatory capacity, faculty readiness, student digital literacy, language context, and available clinical teaching resources. Without such adaptation, AI may widen educational inequities rather than reduce them[
29].
Ethical boundaries and governance frameworks for AI in nursing education
Data privacy and confidentiality
Data privacy is a primary ethical boundary[
11]. Nursing teaching frequently involves real clinical cases, simulation records, reflective assignments, learning analytics, and student performance data[
7]. Students and instructors should not enter identifiable patient information, such as names, admission numbers, medical record screenshots, clinical images, nursing notes, or other identifiable details, into open AI platforms. When real cases are used for teaching, de-identification should be completed before AI-assisted adaptation or discussion. Data minimization should also be applied to student learning data: only data necessary for educational purposes should be collected, and access, storage, secondary use, and deletion procedures should be clearly defined. Privacy protection should therefore be managed at the curriculum and institutional level rather than relying solely on individual teacher judgment.
Bias, equity, and accessibility
AI-generated educational content may reproduce or amplify bias in language, culture, clinical assumptions, and resource availability[
29]. For example, a generated nursing case may implicitly assume access to advanced monitoring devices, sufficient staffing, standardized digital records, or health insurance conditions that are not available in all educational or clinical settings. It may also reflect English-language training data and fail to capture local cultural expectations, minority populations, rural health contexts, or community-based nursing needs. Therefore, AI-generated cases and feedback should be reviewed for cultural relevance, local clinical feasibility, and equity[
11]. Instructors should adapt AI-generated materials to the learners’ clinical environment, available resources, and patient population, especially when teaching in resource-limited or culturally diverse contexts.
Transparency and accountability
Transparency and accountability are essential for responsible AI use in nursing education. Students should know when AI is used to generate learning materials, provide feedback, or support assessment[
30]. Similarly, educators should understand the limitations of AI outputs and avoid presenting generated content as verified professional knowledge[
11]. In formative assessment, AI may provide preliminary feedback or identify learning patterns, but the final interpretation of student competence should remain under human responsibility. The human-in-the-loop principle is especially important in nursing education because competence assessment involves not only knowledge and technical skills, but also ethical sensitivity, communication, professionalism, and patient safety. Institutions should clarify who is accountable for AI-supported teaching materials, feedback, and assessment decisions.
Academic integrity and responsible student use
Academic integrity should be governed through explicit course policies rather than vague restrictions[
22]. Nursing programs should specify which assignments permit AI assistance, which tasks must be completed independently, and when AI use must be disclosed[
16,
27]. For case analyses, nursing care plans, reflective writing, and academic reports, students should be required to document how AI was used, verify generated claims against reliable sources, and provide their own clinical reasoning. AI-generated text should not be accepted as a substitute for students’ professional judgment or independent analysis. A process-oriented approach may be more appropriate than a simple prohibition: students should be taught to compare AI outputs with textbooks, clinical guidelines, and faculty feedback, and to identify errors, missing patient context, unsupported claims, and hallucinated references[
31,
32].
Human oversight and professional responsibility
Drawing on established ethical guidance for AI in health professions education, ethical AI governance in nursing education should follow the principles of transparency, accountability, fairness, privacy protection, beneficence, non-maleficence, and human oversight[
11]. AI can support information organization, formative feedback, case simulation, and learning reflection, but it cannot assume professional responsibility for educational or clinical judgment. Educators remain responsible for curriculum design, content validity, assessment interpretation, and value guidance. Clinical preceptors remain essential for evaluating students’ bedside performance, communication, ethical sensitivity, and patient-safety awareness[
33]. Therefore, AI should be embedded in a supervised educational workflow in which human educators retain final responsibility for teaching decisions, assessment outcomes, and professional formation[
28].
Irreplaceable domains of nursing practice
Certain domains of nursing practice are inherently irreplaceable by AI. Nursing is not only a cognitive or procedural activity but also a relational, embodied, ethical, and accountable practice. AI may simulate communication scripts, generate patient scenarios, and prompt reflection, but it cannot truly perceive patients’ fear, pain, loneliness, family pressure, cultural concerns, or value preferences. It cannot replace bedside observation, hands-on care, clinical empathy, emotional responsiveness, ethical judgment, responsibility in critical situations, interprofessional collaboration, or the professional values transmitted by educators and clinical preceptors. Therefore, AI should be used to strengthen preparation for clinical practice, not to replace authentic nurse-patient relationships or real clinical experience. Nursing education should make this boundary explicit so that students understand AI as a supportive learning tool rather than a substitute for human presence, moral responsibility, and patient-centered care[
33].
Conclusion
AI is becoming an important educational variable in nursing education, but its value depends on responsible integration rather than simple adoption[
6,
7]. When aligned with curriculum objectives and faculty supervision, AI may support personalized learning, case-based teaching, simulation training, formative feedback, and AI literacy development[
6,
28]. However, current evidence remains uneven, and many reported benefits are based on short-term outcomes or perceived competence rather than long-term clinical transfer[
6,
34]. Nursing programs should therefore integrate AI cautiously, using it as a supervised learning scaffold rather than as an authority for clinical or ethical judgment[
21]. Effective implementation requires clear curriculum policies, evidence verification, privacy protection, transparent accountability, academic integrity management, and faculty development[
30]. Most importantly, AI must not replace the human-centered foundations of nursing practice[
21]. Empathy, bedside care, ethical responsibility, professional judgment, and authentic nurse-patient relationships should remain central to nursing education in the digital era[
21].
The Author(s) 2026. This article is published by Higher Education Press at journal.hep.com.cn.
PDF
(247KB)
