Nursing Students' Perspectives on the Use of Artificial Intelligence in Psychiatric Nursing Education: A Qualitative Meta-Synthesis
Article information
Abstract
Purpose
This study aimed to explore nursing students' perspectives on the use of artificial intelligence (AI) in psychiatric nursing education through a qualitative meta-synthesis.
Methods
The qualitative meta-synthesis was conducted from August 13 to August 24, 2025, utilizing searches in the CINAHL, OVID, PubMed, Scopus, and Web of Science databases. The review included qualitative studies published in English that focused on AI applications in psychiatric nursing classroom instruction or practicums among nursing students. The selected studies were analyzed and synthesized following the steps outlined by Thomas and Harden.
Results
Five studies that utilized virtual reality programs as a form of AI in psychiatric nursing practicums were included. Four overarching themes emerged: (1) effective learning tool; (2) enhancement of self-directed learning; (3) innovation in nursing education; and (4) challenges in replicating complex clinical environments.
Conclusion
AI-based tools have the potential to enhance both clinical and theoretical psychiatric nursing education by providing enriched learning experiences tailored to the characteristics of each course. These findings offer valuable insights for developing costumized AI programs that address the unique aspects of psychiatric disorders, thereby supporting the advancement of AI-integrated educational strategies in nursing.
INTRODUCTION
1. Background
Artificial Intelligence (AI) is a groundbreaking form of machine intelligence designed to perform tasks that typically require human abilities, such as thinking, learning, decision-making, and adapting to the environment [1]. In addition to these basic capabilities, AI is defined as technology capable of executing higher-order cognitive functions, including reasoning, planning, and creativity [2]. Particularly in the field of education, AI has been utilized in diverse ways, such as intelligent tutoring systems, virtual simulations, data analysis for learning, and predictive analytics [1].
To adapt to the COVID-19 pandemic, nursing programs shifted to online education, during which AI-based tools were introduced to enhance learning [3]. Integrating AI in nursing education addresses the limitations of traditional teaching methods by enabling individualized learning experiences, customized feedback, and broad access to resources that transcend time and location constraints [4]. Moreover, through automated feedback and assessments, AI not only enriches the learning environment and improves academic performance, but also influences student engagement, evaluation methods, and pedagogical approaches, ultimately strengthening essential nursing competencies, including clinical decision-making and critical thinking [5,6].
This transformation is particularly relevant in psychiatric nursing education, a vital component of the nursing curriculum that demands both theoretical understanding and practical skills to provide effective care for patients with mental illness [3]. Theoretical instruction emphasizes critical thinking, understanding psychiatric disorders, communication skills, and building trust with patients [7], whereas clinical practice focuses primarily on psychosocial rather than physical assessments [8]. Central to this coursework is the application of the nursing process, with rapport building and therapeutic communication serving as essential elements [8,9]. However, nursing students often have limited opportunities to encounter patients with severe mental illness firsthand [10]. This lack of exposure can cause anxiety during clinical psychiatric training, impeding students' acquisition of knowledge, confidence, and clinical skills [11]. To address these challenges, AIbased tools have been proposed as an effective means of integrating theoretical knowledge with experiential learning in psychiatric nursing education [10].
Previous research on the use of AI in psychiatric nursing education has reported notable benefits for nursing students, such as improved communication skills, greater cognitive engagement, enhanced affective empathy, increased learning motivation, and reduced anxiety during patient interactions [3]. Moreover, AI programs have been shown to support repeated learning opportunities, deepen the understanding of mental illness, strengthen therapeutic communication skills, and foster self-efficacy [10]. Despite these promising outcomes, integrating of AI tools within educational and clinical settings remains limited, largely because of insufficient ongoing AI training, a shortage of qualified faculty, and pressing ethical concerns [12]. Therefore, gaining a deeper understanding of nursing students' perceptions of AI use in psychiatric nursing education is of paramount importance as their perspectives are critical for ensuring the successful integration and long-term impact of these innovative educational approaches [9].
To achieve such an in-depth understanding, it is essential to synthesize the fragmented findings of previous qualitative studies because no single study can fully capture the complexity of students' perspectives. Qualitative meta-synthesis is a method that respects the unique contributions of primary studies while generating new insights through the extraction, reconstruction, and integration of key concepts [13].
2. Aim
This study aimed to explore nursing students' perspectives on the use of AI in psychiatric nursing education through a qualitative meta-synthesis.
METHODS
1. Study Design
This qualitative meta-synthesis was conducted to explore nursing students' perspectives on AI use in psychiatric nursing education. Thematic synthesis was employed to synthesize the findings of the included studies [13]. This review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [14].
2. Study Identification
To identify qualitative studies on nursing students' perspectives on AI use in psychiatric nursing education, a systematic search was conducted in the following databases: Cumulative Index to Nursing and Allied Health Literature (CINAHL), OVID, PubMed, Scopus, and Web of Science. A literature search was conducted between August 13 and August 24, 2025, without restrictions on the year of publication. To effectively integrate keywords, the search strategy utilized a combination of terms linked by the "AND" and "OR" operators. The search terms were grouped as follows:
• AI-related terms: "artificial intelligence," "AI," "generative artificial intelligence," "generative AI," "genAI," "GAI," "machine learning," "ML," "deep learning," "natural language processing," "NLP," "decision support systems, clinical," "clinical decision support system," "CDSS," "virtual reality," "VR," "virtual simulation," "virtual*," "large language model," "LLM," "robotics," "chatbot*," "chatGPT," "AI-based*," "AI-assisted," "digital mental health tools"
• Psychiatric nursing terms: "psychiatric nursing," "mental health," "mental health nursing," "psychiatric mental health," "psychiatric mental health nursing," "psychiatric care"
• Psychiatric nursing education and practice-related terms: "psychiatric nursing education," "mental health education," "psychiatric nursing practice," "psychiatric nursing practicum," "psychiatric nursing training," "psychiatric nursing simulation," "psychiatric clinical practice," "psychiatric clinical practicum," "psychiatric clinical training," "mental health simulation," "mental health nursing practice," "mental health nursing practicum," "mental health nursing training," "mental health clinical practice," "mental health clinical practicum," "mental health clinical training"
• Nursing students' terms: "students, nursing," "nursing student*," "student nurse*," "undergraduate nursing student*," "pupil nurse*"
3. Inclusion and Exclusion Criteria
Articles were selected according to the following criteria: 1) qualitative studies that addressed the use of AI in psychiatric nursing classes or practicums among undergraduate nursing students; 2) studies that utilized qualitative research methods, such as individual or focus group interviews; 3) peer-reviewed studies published in English; and 4) studies with full-texts available. The exclusion criteria were as follows: 1) qualitative studies involving graduate nursing students or licensed nurses; 2) studies including mixed samples of undergraduate nursing students, graduate students, and nurses, where data specific to undergraduate students could not be independently extracted; 3) studies employing methodologies other than qualitative research, such as mixed-methods, systematic reviews, or meta-syntheses; 4) studies on e-learning, online lectures, or simulations that do not incorporate AI; 5) studies published in languages other than English; 6) unpublished research, including abstracts, posters, dissertations, or conference proceedings; and 7) studies for which full-text access was unavailable.
4. Study Selection
Two researchers (EP and SS) independently performed the literature search, and any disagreements were addressed through discussion until a consensus was reached. The database search identified 5,762 studies, of which 344 duplicates were removed, leaving 5,418 studies for title and abstract screening. Of these, 4,955 were excluded because they did not employ qualitative research. Of the 463 full-text articles assessed, 458 were excluded for the following reasons: 423 were unrelated to AI, seven included participants other than undergraduate nursing students, 23 studies did not pertain to psychiatric nursing classes or practicums, one was not published in English, two employed quantitative methods, and two reported analyses unrelated to AI. Five studies were ultimately included according to the inclusion and exclusion criteria. The systematic process of study selection is shown in Figure 1.
Flowchart of the search process.
5. Quality Assessment
Two researchers independently conducted a quality assessment of the included studies utilizing the Critical Appraisal Skills Programme (CASP) Qualitative Studies Checklist [15], which comprises 10 questions designed to evaluate the trustworthiness, credibility, and methodological rigor of each study. These criteria address the clarity of the aims, appropriateness of the qualitative methodology and design, adequacy of recruitment and data collection, consideration of researcher-participant relationships, ethical issues, rigor of data analysis, clarity of findings, and the overall value of the study. In the absence of standardized scoring or cutoff criteria, each question was rated as "Yes," "No," or "Can't tell." Discrepancies were resolved by discussion to reach consensus. Following this assessment, all five studies were retained for inclusion in the review.
6. Data Synthesis
This study employed the systematic qualitative synthesis method proposed by Thomas and Harden [14]. The analysis proceeded in three stages. Initially, two researchers (EP and SS) independently reviewed the selected studies, immersed themselves in the content through repeated readings, and categorized quotations based on their meaningful context to ensure a comprehensive and insightful examination. Subsequently, a detailed analysis of the coded data was performed to identify key similarities, differences, and relationships, thereby deriving descriptive themes that shaped the overarching narrative. Finally, these themes were synthesized to capture nursing students' perspectives on the use of AI in psychiatric nursing education, providing new insights beyond the scope of individual studies.
RESULTS
1. Study Characteristics
This qualitative meta-synthesis included five studies, all of which utilized virtual reality (VR) programs as AIbased tools in psychiatric nursing practicums, and were published between 2022 and 2025. The participants were undergraduate nursing students from three academic years: the second (A5), third (A2 and A4), and fourth years (A1 and A3). Across the studies, there were 69 participants in total, including 15 men and 39 women; in one study, the gender of 15 participants was not reported (A5). Two studies employed convenience sampling (A2 and A5), one used purposive sampling (A3), and two did not specify their approach (A1 and A4). The studies were performed in South Korea (A1 and A3), Norway (A4), and the United Kingdom (A5), with one study not reporting the country (A2). Data collection settings also varied; one study used vSim in an online practical training program (A1) and another involved a two-week virtual simulation-based practicum (A3). One study reported hospital placements in a mental health ward for patients with dementia (A2), while another described an eight-week clinical placement that included one day at the university's simulation center, divided between traditional and VR simulations (A4). Finally, one study included a pre-registration program in which participants experienced both AI- and VR-based interaction modes during the the simulated placements (A5). Various data-collection methods were used in the included studies. Two studies utilized non-face-to-face approaches: one involved conducting online focus group interviews (A1), and the other combined observation with semi-structured interviews (A2). One study involved semistructured individual interviews via telephone (A3). The remaining two studies employed in-person focus group interviews (A4 and A5), with the A4 study additionally collecting reflective notes. Three studies applied content analysis (A1, A2, and A4), whereas two studies used thematic analysis (A3 and A5) (Table 1).
Characteristics of Included Studies
2. Quality Assessment of Included Studies
The methodological quality of the included studies was assessed using the CASP Checklist (Table 2). Two researchers independently appraised each study, and any discrepancies were addressed through discussion until an aggrement was reached. All five studies included in this review adhered to the CASP checklist criteria with respect to the clarity of aims, appropriateness of qualitative methodology and design, data analysis, clarity of findings, and overall value of the study. Two studies (A1 and A4) did not specify the sampling method used for participant recruitment. One study (A2) lacked sufficient details about the data collection procedures, including interview setting and timing, and data saturation. Four studies (A1, A2, A4, and A5) did not report on the relationship between the researchers and participants. One study (A4) did not indicate whether Institutional Review Board (IRB) approval was obtained.
Quality Assessment of the Studies
3. Meta-Synthesis
The synthesis of five studies on nursing students' perspectives on the use of AI in psychiatric nursing education identified four key themes: "effective learning tool," "enhancement of self-directed learning," "innovation in nursing education," and "challenges in replicating complex clinical environments" (Table 3).
Synthesized Themes of Nursing Students' Perspectives on the Use of Artificial Intelligence in Psychiatric Nursing Education
1) Effective learning tool
The nursing students recognized the VR program as a powerful educational resource that effectively bridges the gap between theoretical knowledge and clinical practicum. Instead of passively acquiring knowledge, they engaged in a dynamic environment in which theoretical content was reinforced through practical application and professional roles could be explored beyond the constraints of traditional placements.
(1) Facilitating understanding of mental illness
Nursing students, who rarely had opportunities to interact directly with individuals with mental illness, reported that the VR program provided meaningful educational experiences. This allowed them to observe psychiatric symptoms, such as delusions (A1), and systematically collect and analyze related data (A5), thereby facilitating the identification and implementation of appropriate interventions (A5). This process clarified content that had previously been confusing in class, thereby substantially supporting students' learning (A4).
Since I typically do not have opportunities to closely interact with patients with schizophrenia, I did not know exactly how to manage such patients. I learned about schizophrenia from seeing first-hand that patients can refuse to take medication because of delusions of poisoning, and also by performing MSEs and reviewing the contents of conversations. (A1)
I think you get a lot of benefits, especially in observing symptoms and how the nurse approaches the patient. What the nurse does to handle a situation, to de-escalate, I think you get a lot more benefit from that than theory, lectures and taking notes. I am very practical, so I learn much more from things like this and in practice than from a lot of theory that gets mixed up. (A4)
(2) Training opportunities for nursing role performance
Participants perceived the VR program to be highly valuable because it allowed them to experience various nursing roles that were typically challenging or restricted in real clinical settings (A3 and A4). This experience enhanced their understanding of patient care in specific situations, including appropriate nursing actions (A3), communication skills, and relevant assessments (A4). They also reported that the program provided meaningful preparation for their future roles as nurses (A3).
In the vSim, we do everything as nurses; therefore, taking care of a patient completely, that is, beyond what you're allowed to do as a student, by clicking on features is the greatest change. It was helpful because I was able to learn and understand that you should act like this in this case. It was good that I was able to do things that I wasn't allowed to in hospitals and gained an indirect experience with things like how to address these, as a prospective nurse. (A3)
2) Enhancement of self-directed learning
Nursing students considered the VR program a learning space that fostered independence and active engagement. It not only boosted their confidence in applying skills to real clinical situations but also offered the flexibility to learn at an individualized pace, revisit key content, and explore additional resources. Through these experiences, the students gained more control over their learning and enhanced their confidence and readiness for clinical practice.
(1) Building confidence in clinical application
Participants expressed that the VR program significantly bolstered their confidence in applying their clinical skills in real-world settings. Participants noted that the chance to perform nursing interventions, such as medication administration, and check their outcomes provided them with a profound sense of achivement (A3). They developed competencies through repeated, patient assessments and the meticulous gathering of essential information, which apply seamlessly to actual clinical situations (A5). Furthermore, they employed therapeutic communication strategies, which enhanced their skills and instilled a strong sense of confidence in their ability to interact confidently during face-to-face encounters (A5).
Based on patient symptoms, I was able to report to doctors and administer the prescribed medications. Since it's online, I could check the outcomes of my interventions. There are more opportunities in the virtual simulation practicum and passing the scenarios during the first try gave me a sense of achievement. (A3)
I think the AI section, the virtual section has actually given me a very strong sense of belief in myself that I will be able to carry out certain things if I'm asked to do them for the fact that I've actually seen it being done. That has actually increased my competency in most parts of the placement. It's good. (A5)
(2) Flexible approaches to self-directed learning
Nursing students perceived that, unlike clinical practicums, the VR program provided a more flexible learning environment with fewer time constraints, thereby enabling full engagement in the learning process. They could control their pace and revisit the content as needed, which helped them address gaps in their theoretical knowledge, and explore questions that arose during learning (A3). Access to diverse resources allowed for in-depth learning experiences that were difficult to achieve in actual clinical settings (A3).
vSim provided ample time for self-study, which was great for filling in the gaps in my theoretical knowledge. The ability to manage my time was probably the biggest advantage. (A3)
During clinical practicum, I sometimes had to skip questions because I had no time, but in virtual simulation, I felt I could search for anything that I was curious about. (A3)
3) Innovation in nursing education
Nursing students perceived VR-based learning as a future-oriented approach to nursing education that goes beyond lecture-based education. They highlighted that the program provided immediate feedback, helping them recognize and improve their weaknesses; created a sense of safety by reducing anxiety and pressure often present in psychiatric practicums; and offered a highly immersive environment that enhanced concentration and realism in learning. Overall, these aspects positioned VR as an innovative mode of nursing education, effectively bridging the gap between theory and practicum, while better preparing students more effectively for their future professional roles.
(1) Immediate feedback mechanisms
Participants highlighted that the VR program offered an innovative feature of immediate feedback, enabling real-time recognition of performance gaps. All simulation activities were recorded and paired with automated summaries, allowing students to identify overlooked elements or inappropriate responses and make prompt corrections (A1 and A5). This technologically driven process strengthened their clinical knowledge and enhanced the overall quality of the learning experience (A5).
If I had been performing real clinical practice, no matter what I reviewed later, there would always be something regarding my performance I would miss, but with Sim everything I did was recorded and I could immediately obtain feedback. (A1)
It does actually give a very good feedback […] It furnishes your clinical knowledge, physical and mental. (A5)
(2) Safe and supportive learning environment
Participants consistently indicated that the VR program profoundly transformed their learning experience by providing a more comfortable and less intimidating atmosphere. Nursing students valued the opportunity to gain experience in face-to-face encounters with patients before entering clinical practice (A1). They emphasized that within this safe space, they could repeatedly hone their skills without anxiety about being judged by others (A2). Nursing students reported that, during clinical practice, they often felt psychologically burdened when patients repeatedly asked the same questions. However, in the virtual simulation, they experienced a greater sense of psychological stability because patient interactions were simulated rather than real (A3). Most importantly, students noted that the virtual environment enabled them to focus intensely on learning without fear of encountering unpredictable, aggressive situations (A4).
I suppose to test out your skills first online then you're sort of in a safer environment than when you're face to face with somebody. Learning this online is more of a safe environment to get it wrong and where you can get your skills up to scratch really to be able to do it face to face with somebody. (A2)
For us students who have never experienced something like that, seeing an aggressive outburst in VR gives us a sense of security. At least, I feel very insecure when I enter a practice. Especially in mental health, I know nothing about it and by having felt like you have been part of it before, you feel a bit more secure, you know a bit more about what to expect […]. (A4)
(3) Immersive and interactive learning experiences
Participants reported that the VR program provided an immersive learning environment comparable to that of caring for actual patients. This allowed them to apply and review skills such as therapeutic communication and mental status examination (MSE), previously learned only in abstract terms during theoretical classes, within a context that closely resembled real-world practice (A1). They emphasized the realism of the program, noting that tasks such as managing the demands of multiple patients simultaneously (A2) and engaging in conversations as if with real patients (A1) felt like a reproduction of clinical settings. Furthermore, they highlighted that the immersive environment minimized distractions and enhanced concentration during learning (A4).
I learned therapeutic communication in theory class, but the examples given in the textbook were too formal. On the other hand, Sim appeared realistic, so it felt like I was talking to real patients, which really helped me learn therapeutic communication. […] (A1)
You notice that all your senses are much sharper when you have the whole surrounding and everything is sharper. You pick up more details than a film; you don't get the chance to escape; your attention doesn't drift away to your phone or other things. You have just this to watch and then [you] pay attention differently. (A4)
(4) Integration of novel educational approaches
Nursing students viewed VR-based learning as an innovative educational approach that complemented traditional clinical practicums. Participants emphasized that this innovative program was a powerful learning tool, particularly when they approached it with a strong desire to learn (A1). They expressed that, as the world continues to develop and change, such programs should become essential components of the psychiatry nursing curriculum (A5). They also highlighted that this fosters interactive and open learning, particularly among students who may be hesitant to ask questions (A5).
I would say the world is changing. Based on that fact, we have to be catching up every day as the world evolves because the time might come that maybe we are in a different location, the nurse is in a different location, and the patient's in a different location. With that, with the video from him talking and we're responding, there's a lot we can do to be able to diagnose them correctly. So I really think the changing time is going to be what is going to This is the future, actually. (A5)
I think it's a very good innovation. […] (A5)
4) Challenges in replicating complex clinical environments
Nursing students acknowledged the value of the VR program but noted that it could not fully capture the complexity and authenticity of psychiatric clinical practice. They observed that patient presentations were often oversimplified, with limited variability in symptoms and responses. Crucial nonverbal cues and emotional expressions were missing. The lack of unpredictability and dynamic interactions in the simulated ward reduced realism, making it difficult to replicate the tension and challenges faced in actual clinical encounters. Furthermore, the students expressed concerns that the answer-oriented nature of the VR learning process could limit the development of empathy and hinder the transfer of skills from simulation to authentic clinical practice.
(1) Limited representation of psychiatric symptoms
The participants felt that the VR program did not accurately replicate the diverse range of symptoms and behaviors observed in psychiatric patients during real clinical encounters. They mentioned that the virtual patients in the simulations, lacking appropriate speech pace and eye contact, hindered genuine interactions (A1). Consequently, they felt they were engaging with lifeless machines rather than authentic human beings (A1). Additionally, some participants noted that the absence of facial expressions made it challenging to assess the patients' emotional states, and some program responses were deemed inappropriate (A3). They criticized the program as being overly simplistic, noting that patients with different diagnoses often exhibited similar symptoms and behaviors (A3). Furthermore, students indicated that virtual patients could only recognize specific words, which impeded basic communication, leading them to conclude that such simulations could not fully replace real clinical settings (A5).
I felt that when I used Sim, I could determine the correct statements to use during conversations with patients, but I was not able to learn about nuances such as the pace and intonation to use during conversations or the level of eye contact to use with patients. Also, I did not learn what I would feel while talking to a real person. I had the feeling that I was simply completing a task, rather than interacting with a real person. I was also less careful because I knew the patient was not a real person. (A1)
(2) Risks of the emergence of passive learning behaviors
Participants reported that the VR program sometimes encouraged a focus on selecting the "correct" answers rather than deeply engaging with the virtual patients' emotions or developing empathy (A1 and A5). This tendency could lead to passive learning behaviors, where emphasis is placed on completing tasks correctly rather than fostering reflective or critical thinking. Nursing students expressed concerns that this approach might limit meaningful interactions with the content and reduce the development of essential clinical reasoning skills (A5).
I concentrated on selecting the correct responses to questions, so I do not think I felt any emotion. I had a strong motivation to identify the right answers so, for me, whether the answer was correct or wrong was more important than learning empathy. (A1)
I think it's just developing a lazy way of learning. It's not really developing the confidence. It's just like, okay, I have to just answer these questions. It's not really engaging. (A5)
(3) Lack of dynamism and realism in clinical settings
The participants noted that the VR program was limited in capturing the dynamism and realism of actual psychiatric wards. They reported that patients' symptoms remained static, and unexpected situations never arose, even when incorrect interventions were applied (A1). Virtual patients did not move, and opportunities to observe diverse nursing roles were restricted, which the students found disappointing (A1). Scenarios involving violent or unpredictable behaviors were absent (A3), and sudden patient reactions, commonly observed during physical assessments, did not occur (A5). These limitations collectively underscored the fact that VR simulations, although useful, cannot fully replicate the complexity and unpredictability of real psychiatric settings.
While there are few opportunities to gain practical experience in psychiatric wards, it is regrettable that this programme did not portray the overall atmosphere of the hospital, how the systems operate, and other peoples nursing styles. (A1)
[…] I do not think Sim can be an alternative to actual practical training. (A1)
Should I call it the difference in vibrancy? I've been to a psychiatric hospital. There, you could see everything. People were running around, even fighting. However, in the vSim program, it's just words like 'sad', so it wasn't realistic in terms of the psychiatric aspect. It just felt like a machine. (A3)
(4) Concerns regarding clinical applicability
Participants raised concerns that nursing care in real clinical settings might cause mechanical or emotional detachment, reflecting their interactions with virtual patients (A1). They also questioned whether the skills acquired through the VR simulation could be reliably transferred to real clinical practice because of the significant differences between virtual models and actual patients (A3).
I believe that I must receive emotional training to have a long career as a nurse; because Sim does not have this capability: you might become inclined to treat patients as those you met in virtual reality. (A1)
I haven't actually seen or done this in a hospital, and models are so different from real people. I wondered if I would be able to do well in real clinical settings. (A3)
DISCUSSION
This qualitative meta-synthesis analyzed nursing students' perspectives on the use of AI in psychiatric nursing education. All five studies exclusively applied AI to practicum education, primarily through VR programs, with little attention paid to classroom instruction. The findings indicated that AI-based practicum supports therapeutic communication, enhances problem-solving, reduces anxiety toward patients with mental illness, and improves emotional empathy [3,8], suggesting its potential to enrich theoretical education. The adoption of AI was accelerated during the COVID-19 pandemic when psychiatric clinical practicums in closed wards became restricted [3,16]. Although VR was the dominant tool, recent advances have introduced generative AI, learning management systems, augmented reality, and simulation-based platforms, which continue to demonstrate innovative potential in nursing education [2,11,17]. However, most interventions were short-term, ranging from half-day to 10-day sessions, limiting the evaluation of long-term effects on nursing competence [18].
The quality appraisal of the included studies revealed insufficient reporting of participant recruitment, data collection, and ethical considerations. In particular, four studies did not address the researcher-participant relationship. Future qualitative research should strengthen its transparency and rigor by explicitly addressing these issues. The following discussion integrates these findings to highlight the educational potential and limitations of AI in psychiatric nursing education, from nursing students' perspectives.
First, AI-based VR programs offer nursing students unique opportunities to engage with psychiatric symptoms and clinical scenarios rarely encountered in everyday practice, thereby deepening their understanding of mental illness. By providing a controlled environment in which students can directly apply the nursing process and practice therapeutic communication, VR strengthens the connection between theoretical learning and practical application, and fosters cognitive and professional development. Immersive experiences support students in identifying nursing problems, making informed decisions regarding interventions, and refining essential clinical skills that are difficult to achieve in traditional classrooms or limited clinical practicum settings [8,10]. Consequently, VR simulations represent a valuable educational tool that complements conventional teaching methods and can be strategically integrated into psychiatric nursing curricula to maximize learning outcomes.
Additionally, VR programs allow students to experience diverse nursing roles that are typically restricted during real-world clinical practicums. Exposure to multiple professional responsibilities, even in a simulated environment, enhances students' competencies by familiarizing them with key aspects of nurse-patient communication, clinical judgment, and decision-making processes. Similarly, prior research has indicated that VR simulations developed from real clinical cases provide a sense of realism and facilitate indirect engagement with nursing tasks across different settings, including psychiatric and obstetric wards, thereby supporting active and independent learning [11,19]. By offering enriched and interactive experiences, VR enables students to safely practice complex interventions, strengthen their professional confidence, and prepare for future clinical responsibilities.
Second, participation in the VR simulations strengthened nursing students' confidence in clinical applications by providing opportunities to engage with psychiatric cases before entering practice. Similarly, previous studies have confirmed that VR-based training promotes greater improvement in self-efficacy than traditional simulations [8,20]. Experiencing mental illness scenarios in a safe and controlled environment allowed students to acquire essential communication and clinical skills, thereby fostering confidence in their ability to interact effectively with psychiatric patients [3,10]. This suggests that VR functions as a preparatory platform equipping students with competence and assurance as they transition to real-world practice.
VR programs also promoted self-directed learning by enabling students to manage their learning at their own pace without temporal or spatial constraints. The flexibility of VR allowed learners to access immersive simulations anytime and anywhere, increasing motivation and engagement while overcoming the limitations of traditional laboratories [20,21]. Importantly, the option to repeat complex scenarios according to individual needs and preferences provided personalized learning opportunities, thereby enhancing competence and autonomy [19]. This aligns with broader evidence that AI-enhanced education fosters active exploration and deeper engagement with knowledge, positioning VR as an effective tool for cultivating lifelong learning skills among nursing students [21].
Third, nursing students perceived the VR programs positively, as they provided immediate feedback, offered a safe and risk-free learning environment, and enhanced realism through a high level of immersion. Simple feedback mechanisms, such as displaying scores or indicating whether learning goals were achieved, encouraged repeated practice and promoted self-directed learning, whereas overly detailed feedback was reported to cause learner fatigue and reduce engagement [20,22]. Practicing in a virtual environment without direct interaction with real patients allowed students to make mistakes without negative consequences and train in clinical skills and decision-making in realistic yet safe scenarios [20,23]. Particularly in challenging situations, such as managing aggressive patients, VR simulations were considered less threatening than traditional face-to-face practice [9,20,22]. Students who reported fear or anxiety about interacting with patients with mental disorders also valued VR as providing a safe space to practice communication and reduce their apprehension toward clinical practice [22].
A key advantage of VR learning was the immersive experience. Students described a strong sense of presence and perceived interactions with virtual patients and healthcare providers as realistic, which contributed to the development of clinical reasoning skills [19,20,22]. High levels of realism not only increased learner satisfaction [24] but also enhanced motivation and fostered emotional engagement with virtual patients [3]. AI-based simulations strengthened learning outcomes by creating interactive and immersive environments that closely resembled real clinical settings [25].
Taken together, these findings suggest that VR is not only a supplementary tool but also an innovative educational approach. By exposing students to situations rarely addressed in traditional curricula, VR provided novel and meaningful learning opportunities [9,26]. In particular, VR scenarios related to schizophrenia were perceived as easy to use, engaging, and pedagogically valuable [22]. Furthermore, by reflecting on the uncertainty and complexity of real clinical environments, VR simulations serve as strategic tools to foster core competencies in clinical reasoning and decision-making among nursing students [19].
Despite these advantages, nursing students identified notable challenges in replicating authentic clinical environments. Virtual patients often displayed limited symptom variability, repetitive behaviors, and insufficient nonverbal cues, which hindered genuine interactions and the cultivation of empathy [19,20]. The lack of unpredictability and restricted interactivity raised concerns about the transferability of VR-acquired skills to real-world practice. Students also reported that the program sometimes encouraged passive learning behaviors, with a focus on selecting correct answers rather than meaningfully engaging with patients. These limitations emphasize that VR should be integrated as a supplementary tool to traditional clinical training, ideally incorporating scenario-based exercises that address emergencies, complex patient behaviors, and the full spectrum of nursing responsibilities in psychiatric wards [3,21,24]. Carefully designed VR scenarios that reflect authentic clinical complexity are essential for maximizing educational effectiveness and ensuring that learners are well-prepared for the challenges of real-world practice.
This qualitative meta-synthesis reviewed five studies that exclusively examined the use of VR programs in psychiatric nursing practicums. Although this focus provides valuable insights into immersive learning experiences, it limits our understanding of how AI can enhance theoretical education. Future research should broaden the scope to include both practical and theoretical learning, and explore diverse AI-based educational tools beyond VR. Despite this limitation, this study makes a significant contribution by synthesizing nursing students' perspectives on AI, particularly VR-based learning, thereby demonstrating the educational potential and practical value of VR in psychiatric nursing education and providing foundational insights to guide the development of future AIsupported instructional strategies.
CONCLUSION
Nursing students recognize AI as a beneficial tool in psychiatric nursing education because it provides learning opportunities, even without direct patient encounters, and allows engagement with diverse nursing roles. AI supports self-directed learning without temporal or spatial constraints; enhances students' confidence in clinical applications; delivers immediate feedback; and creates safe, immersive, and realistic practice environments that foster clinical competence. Nevertheless, limitations remain, including restricted expression and interaction of psychiatric symptoms, insufficient realism, emergence of passive learning behaviors, and uncertainty regarding transferability to real-world practice. These findings suggest that AI should complement rather than replace traditional educational methods, and underscore the need for tailored, validated AI programs that reflect the unique characteristics of psychiatric disorders.
Notes
Eunyoung Park has served on the journal’s editorial board since January 2018 but was not involved in the decision to publish this manuscript; no other conflicts of interest were reported.
AUTHOR CONTRIBUTIONS
Conceptualization or/and Methodology: Park, E
Data curation or/and Analysis: Son, S & Park, E
Funding acquisition: Park, E
Investigation: Son, S & Park, E
Project administration or/and Supervision: Park, E
Resources or/and Software: Son, S & Park, E
Validation: Son, S & Park, E
Visualization: Son, S & Park, E
Writing: original draft or/and review & editing: Son, S & Park, E