Content area
Aim
This study aimed to systematically review and synthesize the most recent qualitative studies on frontline nurses' insights and perspectives regarding the use of artificial intelligence (AI) tools in their clinical practice in hospital settings.
Background
There is limited information on frontline nurses' perceptions, attitudes and expectations regarding the adoption of AI in healthcare.
Design
A systematic review and thematic synthesis of qualitative evidence was conducted.
Methods
A systematic search was conducted across five electronic databases—CINAHL, PubMed, Web of Science, Ovid and Science Direct—to identify qualitative studies published between January 2020 and December 2024. After selecting studies, a thematic synthesis was performed. This review followed the PRISMA checklist, was registered with PROSPERO and included a quality appraisal of the retrieved studies.
Results
Nine qualitative studies published between 2022 and 2024 were included in this systematic review. The included studies were conducted in five countries at university-affiliated or tertiary hospitals. Participants included 140 frontline nurses and nurse managers with prior experience using technology tools in clinical settings. Five common themes were identified: ethical issues; increased workload; seamless and efficient patient care; reinforcement rather than replacement; and AI as a future nursing care solution
Conclusions
The five themes identified in this review provide valuable insights into how AI tools can be integrated into current and future frontline nursing practices to enhance patient care. Nurse leaders and healthcare policymakers can use these findings to improve nursing research, facilitate the adoption of new AI tools and support their implementation in healthcare settings.
1 Introduction
Artificial intelligence (AI)-based technologies have been widely applied and integrated into global healthcare in recent years ( Pailaha, 2023). AI has significantly transformed and revolutionized healthcare systems. While AI tools have been rapidly studied and incorporated into healthcare workforces, they primarily aim to enhance human well-being and improve patient care ( Douthit, 2022). AI is driving a paradigm shift in healthcare by facilitating accurate clinical diagnoses, supporting decision-making and enabling the analysis of large volumes of data and information ( Pailaha, 2023; Pancho, 2024; Rony et al., 2024a). Although there is no universally accepted definition of AI in healthcare, it is commonly understood as "the ability of computers to independently convert data into knowledge to guide decisions or autonomous actions" ( Douthit et al., 2022) and "the ability of computers to perform tasks associated with humans" ( Oke & Cavus, 2024).
AI adoption and engagement are also inevitable in nursing and their implementation is reshaping nursing practice ( Berlin and Murphy, 2024). Traditionally, nurses and nurse practitioners (NPs)—the world’s largest healthcare workforce—have played a pivotal role in healthcare delivery. They provide bedside clinical treatments and offer holistic care to support patients' well-being ( Flarey, 2024; Sher, 2024). The introduction of AI tools into nursing practice has significantly transformed the roles and responsibilities of nurses. Moreover, the limited source of a theoretical framework elucidating the relationship between AI and nursing practice is significant in both practical and theoretical aspects of nursing.
According to a recent national survey conducted by McKinsey and the American Nurses Foundation in 2023 ( Berlin and Murphy, 2024), 64 % of nurses expressed a willingness to incorporate AI into their practice, while 42 % believed that AI would improve patients' quality of care. However, 23 % of surveyed nurses reported concerns about AI integration, particularly regarding its possible risks to patients' safety ( Berlin and Murphy, 2024). There have been diverse opinions on AI's potential benefits and risks in nursing ( Wieben et al.,2024). Despite these discussions, a deep understanding of frontline nurses' attitudes and insights towards AI remains limited. Exploring nurses' perspectives is crucial to ensure AI acceptability and facilitate effective collaboration between nurses and AI tools. Therefore, this study aimed to conduct a systematic literature review and thematic synthesis of qualitative studies to identify common issues surrounding the adoption of AI in nursing care.
2 The review
2.1 Aim
This qualitative systematic review aimed to identify qualitative studies published in the past five years that examined frontline nurses' perceptions of using AI tools in future healthcare settings. Additionally, it sought to synthesize common themes related to integrating AI into nursing practice. The review addressed two research questions: (1) What are the perspectives of nurses using AI tools in patient care in the near future? (2) What challenges do frontline nurses face when using AI tools in their clinical practice?
2.2 Design
This study employed a qualitative systematic review methodology with a synthesis of qualitative evidence ( Butler et al., 2016; Thomas and Harden, 2008). Using the qualitative thematic analysis method, this review synthesized common themes ( Thomas and Harden, 2008). It adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines ( Page et al., 2021) and followed the ENTREQ statement—"Enhancing transparency in reporting the synthesis of qualitative research"—for rigorous reporting ( Tong et al., 2012). This study was registered in the International Prospective Register of Systematic Reviews (PROSPERO) and a quality appraisal was conducted using the Critical Appraisal Skills Program (CASP) 2024. Since no human participants were involved in this study, ethical approval from the authors' affiliated institutions was not required.
2.3 Information sources and search strategy
A comprehensive and systematic search of five electronic bibliographic databases— CINAHL, PubMed, Web of Science, Ovid and Science Direct—was conducted from inception to November 1, 2024. The search identified empirical and primary studies published between January 2020 and December 2024, including online publications. An initial search in PubMed was followed by searches in the remaining four databases. The search terms used were as follows: (nurs* AND [artificial intelligence OR technology OR digital technology OR digitalization] AND [interview OR perspective OR qualitative]). The first and second authors independently screened the titles and abstracts. To enhance rigor and minimize missing data, a specialist librarian reviewed the entire search process. A final confirmation search was conducted on January 1, 2025.
2.4 Eligibility criteria
The inclusion criteria for this review were as follows: (1) empirical and primary studies that employed qualitative designs or approaches; (2) studies that explored the perspective of nurses and aligned with the objectives of this review; (3) participants in the studies were nurses or NPs (caring for adult patients ≥18 years old) who were currently working in healthcare settings such as hospitals or clinics and were interviewed about their perspectives on technology and AI-powered tools in patient care; (4) studies that included narrative descriptions of nurse interviews or synthesized themes from the nurses' viewpoints; and (5) articles published in English in peer-reviewed journals between January 2020 and December 2024.
Studies that included healthcare providers or technicians as participants were considered eligible if most of the sample comprised nurses and their interviews could be clearly distinguished from other participants.
In contrast, studies were excluded if they: (1) focused on nurse regulators who were not working in hospital- or clinic-based settings; (2) primarily involved non-nurse healthcare providers as the main participants; and (3) did not align with the aim of this review.
In this review, not only qualitative studies but also mixed-method studies that reported qualitative data or interview extracts as data were included. However, secondary studies such as reviews, commentary, or editorials without empirical findings were excluded.
2.5 Study selection process and quality assessment
For screening, five electronic database search records were input into EndNote 21 (Clarivate Analytics, Philadelphia, USA), a reference management tool. After removing duplicates, 169 articles remained for initial screening. Of these, 123 articles were excluded based on their titles, leaving 46 articles for abstract evaluation. Ultimately, only nine studies were selected for the final full-text review.
A quality appraisal was conducted using the CASP 2024 Checklist for qualitative research (CASP, 2024). This checklist consists of three sections with ten questions assessing various aspects, including the study's aim, methodological rigor, appropriateness of the study design, recruitment strategy, data collection, data analysis, ethical considerations, clarity of results and overall value of research. During critical appraisal, each of the ten questions was answered with "yes," "no," or "can't tell" if a qualitative study did not transparently report certain elements.
2.6 Data extraction and synthesis
The research team, consisting of all three authors, performed data extraction and synthesis to ensure methodological rigor. For data extraction, the first and second authors created and completed an extraction table. The table included the following: author(s), publication year, clinical setting, study aim, participants, study design, data collection methods, analysis and synthesis methods and study results.
Data synthesis was followed by Thomas and Harden's (2008) three-stage approach to qualitative thematic synthesis. The research team conducted the following steps to extract codes and themes from the nine selected studies: First, the studies were analyzed and recurring codes were identified through manual line-by-line coding by the first and second authors. Second, common and similar codes were grouped into descriptive themes. Lastly, the themes were further organized into higher-order themes to categorize barriers and facilitators relevant to the review questions. The last author supervised, validated and verified this process to ensure trustworthiness and methodological rigor. Finally, common themes were revealed from thematic synthesis.
3 Results
3.1 Study characteristics
Overall, nine studies that met the aim and inclusion criteria of this qualitative systematic review were included: Hassan, El-Ashry, (2024), Helman et al., (2021), King et al. (2023), Laukka et al. (2022), Rony et al. (2024a), Rony et al. (2024b), Uymaz et al. (2024), Wieben et al. (2024) and Zhai et al. (2022).
A total of 150 frontline nurses, including NPs, participated in the studies. The number of participants in the studies ranged from 5–37 nurses. All nine studies reported that their participants had prior experiences using technologies in their nursing practice, such as Electronic Health Records (EHRs) and Clinical Decision Support Systems (CDSS). The participants' ages ranged from their 20 s to 60 s; however, none had clearly specified ethnicities, genders, or educational levels. All participants were recruited using purposive or convenience sampling.
Data were collected through semi-structured or in-depth individual interviews via an online meeting platform (e.g., WebEx®) or face-to-face interactions. Seven studies used thematic analysis, while one employed content analysis ( Laukka et al., 2022). One study interviewed five nurses individually but did not perform data analysis or synthesis; instead, it presented the interviews narratively ( Uymaz et al., 2024).
The research team used the CASP 2024 tool to evaluate the quality of the selected studies. The first and second authors independently assessed the studies and the third author compared the evaluation results. No discrepancies were found among the three authors ( Supplementary Table A). All nine qualitative studies demonstrated methodological rigor, though each had one to three instances of "no" responses on the checklist.
3.2 Main findings
Five common themes were revealed. Two themes were identified as barriers to adopting AI tools in future health care: (1) ethical issues and (2) increased workload. Three themes were identified as facilitators: (3) seamless and efficient patient care, (4) reinforcement rather than replacement and (5) AI as a future nursing care solution (see
3.2.1 Theme 1: Ethical issues
In seven studies, nurse participants expressed concerns regarding several ethical issues associated with applying AI tools in clinical practice ( Hassan, El-Ashry, 2024; Helman et al., 2021; Laukka et al., 2022; Rony et al., 2024a; Rony et al., 2024b; Wieben et al., 2024; Zhai et al., 2022). The most critical issue was patients' data security and privacy. Many studies emphasized the necessity of securing patients' private health information and ensuring data security when using AI tools ( Rony et al., 2024a; Rony et al., 2024b). In Rony et al. (2024a), one participant highlighted the importance of ethical responsibility in AI use: "As AI enters our domain, critical ethical questions arise, particularly concerning patient privacy and data security" (p. 8).
Another ethical concern was the lack of clear legal boundaries and ethical standards for integrating AI tools into nursing practice ( Rony et al., 2024a; Rony et al., 2024b; Zhai et al., 2022). Concerning this issue, one participant in Rony et al. (2024a) stated, "The use of AI requires a delicate balance between innovation and upholding legal and ethical standards" (p. 8). Similarly, nurse participants advocated for standardized and explicit guidelines to regulate AI in healthcare ( Rony et al., 2024b).
Some nurse participants also argued that while AI tools expedite clinical decision-making, they might inhibit nurses' critical thinking ( Hassan, El-Ashry, 2024; Helman et al., 2021; Laukka et al., 2022; Rony et al., 2024b; Wieben et al., 2024). This concern was especially relevant for newly hired novice nurses, who have limited abilities as nurse professionals and may not have the opportunity to develop critical thinking ( Helman et al., 2021).
Two studies reported concerns regarding AI-driven health inequity or unfairness ( Hassan, El-Ashry, 2024; Rony et al., 2024b). AI systems can inherit bias from the data they are trained on, potentially leading to unfair and unequal care for specific patient groups ( Hassan, El-Ashry, 2024). In Rony et al. (2024b), NPs expressed concerns that AI tools might exacerbate healthcare disparities among vulnerable populations due to unequal access to healthcare.
3.2.2 Theme 2: Increased workload
In all nine studies, most nurses expressed stress and concern about acquiring competency in using new AI tools, the additional time required for integration and the increased workload in their clinical activities. Nurses across all studies highlighted challenges related to both education and training.
Participants demonstrated that they needed time to learn and get accustomed to current telehealth tools such as EHRs, which are already overwhelming: "…There's always going to be those people that perhaps are already overwhelmed with the technology. So, they may not be as receptive to this because this is another task" ( Helman et al., 2021; p. 5). Likewise, several frontline nurses reported that using AI tools would require additional education and technical training ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Uymaz et al., 2024; Zhai et al., 2022). For instance, one nurse compared AI training to obtaining a driver's license, stating, "Before the use of AI, training must be given, just like when getting a driver's license" ( Uymaz et al., 2024, p. 12).
Nurses acknowledged they needed extra or additional educational hours learning new AI-driven systems and tools. One nurse in Rony et al. (2024b) emphasized this challenge: "Embracing AI as nursing professionals requires an ongoing commitment to learning" ( Rony et al., 2024a, p. 8). Nurses are already dedicating their full work hours to patient care ( Zhai et al., 2022). In Zhai et al. (2022), nurses expressed feeling overwhelmed by the additional responsibilities: "If the workload is light, I can certainly complete it, but when I have to manage 12 patients on my shift, I really don't feel I will be able to do it (new AI tool) in detail for every patient." (p. 4). One nurse emphasized the importance of proper training: "Training is vital to maximize the benefits AI can offer in our clinical work" ( Rony et al., 2024a, p. 13). Some nurses are worried that dedicating time to learning AI tools might reduce the time available for hands-on patient care ( King et al., 2023; Uymaz et al., 2024; Wieben et al., 2024).
3.2.3 Theme 3: Seamless & efficient patient care
In eight studies, participants stressed that AI tools enhance seamless and efficient patient care in nursing practice ( Hassan, El-Ashry, 2024; Helman et al., 2021; King et al., 2023; Laukka et al., 2022; Rony et al., 2024a; Rony et al., 2024b; Uymaz et al., 2024; Wieben et al., 2024). Many studies highlighted that AI tools significantly reduce clinical errors and predict risks or complications by analyzing large volumes of patient data. Consequently, AI facilitates seamless, preventive and safety-oriented care ( Hassan, El-Ashry, 2024; Helman et al., 2021; King et al., 2023; Laukka et al., 2022; Rony et al., 2024a; Uymaz et al., 2024; Wieben et al., 2024). Additionally, because AI monitors patients continuously, it enables early detection of complications and timely interventions. By integrating AI into patient data analysis, nurses can predict adverse events and mitigate risks ( Laukka et al., 2022).
For instance, one nurse in Helman et al., (2021) stated, "It's not just about what's on the monitor and … in a computer, it's also looking at the patient as well, so tie all this together. It helps me … okay, this patient is unstable. Keep a close eye on him, prioritize your patients based on their acuity of issues" (p. 5). Another nurse described AI tools as a "second set of eyes" ( Wieben et al., 2024, p. 5).
Some participants noted that AI tools improve healthcare efficiency ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024b). They agreed that AI tools support traditional nursing tasks, such as administrative duties and transform routine workflows ( Hassan, El-Ashry, 2024; Laukka et al., 2022). AI tools enable nurses to dedicate more time to patient care by automating routine tasks ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024b). For example, one ICU nurse stated, "AI enhances our (nurses’) expertise. It analyzes data, flags issues early and frees me to focus on patient care"( Hassan, El-Ashry, 2024, 2024, p. 5). Similarly, one NP emphasized AI's efficiency benefits, stating, "Efficiency is crucial for healthcare delivery and AI delivers. It reduces time spent on paperwork, helping us see more patients without compromising care quality" ( Rony et al., 2024b, p. 12). Likewise, AI tools have facilitated frontline nurses' tasks, contributing positively to patient care and outcomes.
3.2.4 Theme 4: Reinforcement rather than replacement
In the included studies, nurses expressed strong trust in their traditional nursing practices, such as providing empathy and human touch. They emphasized that AI should optimize rather than replace real clinicians ( Laukka et al., 2022; Rony et al., 2024a; Rony et al., 2024b; Wieben et al., 2024). A few nurses raised concerns that AI might replace human clinicians or contribute to the dehumanization of care ( Rony et al., 2024b). Participants emphasized that AI cannot replace nurses' roles ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024a; Rony et al., 2024b; Wieben et al., 2024). While nurses rely on technology, they believe their unique value lies in their clinical experiences and intuition ( Rony et al., 2024a; Wieben et al., 2024). One participant stated, "I think AI cannot replace clinicians. It is still supplementary because humans are still needed for interpretation" ( Laukka et al., 2022, p. 3843). Similarly, Rony et al. (2024b) reported that NPs had strong confidence in their practice: "AI can support us, but it can't replace the empathy and understanding we provide. That's what truly makes patient care special" (p. 11). Nurses further asserted, "A meaningful patient-nurse relationship goes beyond technology. While AI offers data-driven insights, our capacity to deliver compassionate care remains unparalleled and central to our role" ( Rony et al., 2024a, p. 9).
Likewise, while AI has transformed certain aspects of nursing work and contributed to patient-centered care and positive outcomes, it cannot replace the human connection between nurses and patients. Instead, AI is seen as a tool that strengthens and enhances traditional nursing capabilities rather than as a substitute ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024a).
3.2.5 Theme 5: AI as the future of nursing care solutions
In all nine studies, participants were convinced that AI tools represent innovative technologies and a paradigm shift in nursing practice, improving patient care coordination. Many participants emphasized that adapting to and using AI and technology in patient care is an essential and unavoidable trend in modern healthcare practices ( Hassan, El-Ashry, 2024; King et al., 2023; Laukka et al., 2022). One senior nurse stated, "Anticipating and embracing AI's potential positions us as pioneers in shaping nursing's future" ( Rony et al., 2024a, p. 10).
The primary advantage of AI adoption is its ability to facilitate and transform nursing tasks. Moreover, some participants viewed AI tools as trusted partners and co-workers ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024b). For example, one NP said, "We see AI as a valuable tool. It's like a trusted colleague who can support us" (Rony et al., 2024b). Likewise, participants expressed willingness to adopt AI, recognizing its essential role in nursing care evolution (Rony et al., 2024a). In these studies, nurses reported being convinced to integrate AI tools into their practice ( Hassan, El-Ashry, 2024; Wieben et al., 2024). Ultimately, they are willing to work alongside AI to improve nursing care delivery.
4 Discussion
This systematic review and thematic synthesis of qualitative studies explored frontline nurses' insights regarding AI tools adoption in clinical practice. Through the analysis and synthesis of data from multiple studies, common themes of AI tools emerged from frontline nurses who work in hospitals. Five overarching themes-categorized as barriers and facilitators-were identified when using AI tools in nursing care.
Two major barriers to AI adoption included (1) ethical issues and (2) increased workload. Meanwhile, the three facilitator themes of adopting AI tools were (3) seamless and efficient patient care, (4) reinforcement rather than replacement and (5) AI as a future nursing care solution. Research on AI and nursing practice has been ongoing for a decade. This qualitative systematic review of facilitators and barriers from the perspective of the frontline nurses may bridge gaps or highlight overlooked aspects that need further acknowledgment.
While nurses agreed that AI tools are promising and continue to evolve in current and future healthcare delivery, they expressed significant ethical concerns. Nurses from multiple studies indicate that the most pressing ethical challenges in AI deployment involve patient data privacy and security threats ( Hassan, El-Ashry, 2024; Rony et al., 2024a). In addition, there remains a lack of standardized guidelines and regulations governing AI integration into current healthcare systems. These ethical issues extend beyond nursing and are recognized as broader medical and global health concerns ( Dunlap, Michalowski, 2024; Marques et al., 2024). Given the priority issues of data security and authorized access, numerous research initiatives and healthcare forums have actively debated potential solutions for the successful implementation of AI in clinical practice ( Shaw et al., 2024; Sung, 2023). Without transparent fundamental ethical frameworks guiding AI development in nursing, it may have detrimental consequences for patient outcomes and global health ( Amann et al., 2020; Rony et al., 2024a). Therefore, it is paramount to strengthen AI-related healthcare ethics in nursing.
This review also identified one of the imperative ethical issues that previous nursing research has not adequately highlighted: health inequity and unfair patient care ( Hassan, El-Ashry, 2024; Rony et al., 2024b). The findings suggest that AI may contribute to health disparities due to biased datasets, AI algorithms, unequal healthcare access, or variations in patients' backgrounds. Moreover, AI tools can potentially exacerbate health disparities ( McCormack, 2024; Rony et al., 2024b). Recent research has reported that AI could negatively affect health equity through biased data and algorithms, geo-economic distribution, race, or gender ( Green et al., 2024; Yang et al., 2024). Similarly, AI tools pose potential risks that may hinder the achievement of equitable healthcare. More research using diverse data and the development of ethical regulations are needed to address these challenges and mitigate health disparities while ensuring fair AI implementation.
As reported in this review, nurses perceived AI tools as increasing their workload. When AI tools are introduced, it is evident that nurses, nurse managers and nurse leaders must learn new technologies and require ongoing technical training ( Hassan and El-Ashry, 2024). The need for education and training was frequently cited as a concern and challenge for nurses. Similarly, quantitative studies have reported that stress levels among nurses significantly increase when they are required to acquire new AI-related technical skills to maintain professional competency ( Alnawafleh, 2024).
This review identified paradoxical findings regarding AI tools. Some nurses perceived AI as increasing their workload and being time-consuming. In contrast, other nurses viewed it as enabling more efficient patient care by reducing time-intensive tasks, thus allowing them to spend more time with inpatients ( Hassan and El-Ashry, 2024). Nurses who see the advantages of AI emphasize that it provides more opportunities for meaningful human connections with patients, a key aspect of nursing care ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024a). Therefore, balancing AI technologies and nurses' patient care responsibilities is essential.
As this review found, nurses generally had a positive outlook on the potential for AI to facilitate seamless patient care. These findings are consistent with research in other healthcare disciplines ( Ali et al., 2023). Previous studies have also demonstrated that AI tools enhance efficiency and timeliness in real-world clinical settings, improving physicians' workflow and systematic reviews of clinical care ( Ali et al., 2023; Wenderott et al., 2024). Such benefits may ease nurses' adaptations to AI-driven practices.
One of the novel findings of this review is that most nurses valued AI tools as co-workers and welcomed their role in supporting and enhancing nursing practices ( Hassan, El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024a). Additionally, this review highlighted that nurses strongly believe AI or any technological advancements cannot substitute or replace the nursing profession because nurses possess unique clinical intuition for patient care, holistic caregiving approaches and empathetic care ( Laukka et al., 2022; Rony et al., 2024a).
Furthermore, many nurses in this review expressed a strong willingness to participate in the early stages of AI tool development ( Hassan and El-Ashry, 2024; Helman et al., 2021; Wieben et al., 2024). Their primary motivation for involvement was to ensure the development of user-friendly AI interfaces tailored to nurses' needs ( Hassan, El-Ashry, 2024, p. 6). Nurses emphasized the importance of collaborating with AI specialists and technicians to design AI tools that seamlessly integrate into clinical workflows ( Helman et al., 2021; Wieben et al., 2024). Therefore, interdisciplinary collaboration, cross-sector partnerships and engagement with technical staff are crucial to ensure optimized AI-assisted nursing care.
This review also identified several gaps in the existing AI and nursing literature. Since the recent evolution of AI in nursing research and practice, studies exploring its theoretical features remain limited ( Wynn et al., 2023) and few studies have clearly conceptualized the intersection between AI and nursing ( Shang, 2021). Several studies included in this review adopted existing theoretical frameworks to explain AI's role in nursing practice ( Hassan, El-Ashry, 2024; Uymaz et al., 2024; Wieben et al., 2024; Zhai et al., 2022). However, none of the nine qualitative studies reviewed generated conceptual models linking AI to nursing care. The synthesized themes identified in this review may serve as a fundamental understanding of AI-based nursing care practices, offering insights from frontline nurses to improve patient outcomes.
4.1 Strengths and limitations
This review has some limitations. First, the included studies did not specify the demographic characteristics of nurse participants, such as their educational level, age, years of experience as nurses, or competency in using technology. Notably, perspectives on technologies and AI tools may differ between novice and experienced nurses. Second, this review included qualitative studies that broadly explored the nurses' perceptions of AI and general issues surrounding AI integration in nursing. However, instead of examining specific AI-equipped tools used in clinical applications, the included studies primarily focused on the broader outlook of AI tools in future nursing practice. Only the study by Helman et al., (2021)) explored nurses' perspectives on a Graphical User Interface (GUI), which integrates AI for inpatient risk prediction and monitoring. Third, although all the studies were published recently, most interviews were conducted in only six countries. Each country has its unique healthcare system, financial and government support for AI systems and policies for adopting new technology in government-run healthcare. However, the thematic similarities across studies suggest that frontline clinical nurses share common perspectives on AI tools. Fourth, this review selected only studies published in English, which means that findings published in other languages were excluded. Furthermore, the included studies focused on frontline nurses working in hospitals who directly care for inpatients. Finally, this study could not elucidate a theoretical framework for the relationship between AI and holistic nursing practice, which would hold both practical and theoretical significance in nursing.
Despite these limitations, this review has several strengths and makes new contributions to evidence-based knowledge. This is the first study to synthesize frontline nurses' perspectives on AI adoption and use in their clinical practice. The identified themes provide valuable insights that can inform the development and modification of current and future AI collaborations in nursing.
4.2 Implications for nursing research and practice
Nurse leaders can use the themes identified in this review to shape future nursing care environments, incorporating AI tools for patient care, professional development and healthcare system improvement. However, this review focused only on hospital-based frontline nurses; therefore, future research should explore AI adoption in other healthcare settings, such as community care and nursing homes.
Compared with other healthcare practitioners, nurses have been relatively slow to adapt to AI in clinical practice ( Douthit et al., 2022; Shang, 2021; Shi et al., 2023). Additional rigorous empirical studies are required to explore the impact of AI-driven nursing interventions on patient outcomes. As more research is conducted, evidence-based AI integration in nursing can be further refined and expanded.
This qualitative synthesis highlights that AI cannot replace nurses in patient care, as nurses play a unique role in providing advocacy and holistic support for patients. Therefore, healthcare policymakers and nursing leaders must recognize the value of nurses in the AI-driven healthcare era and invest in additional nursing education, research and continuous support for nursing practices. Moreover, nurses’ responses to AI vary according to their role definitions and relationship with technology ( Douthit et al., 2022). Therefore, training programs should be tiered according to these differing needs.
5 Conclusion
This study is the first systematic review to thematically synthesize recent literature on frontline nurses' perceptions and insights on AI adoption in hospitals. Nurses and other healthcare professionals can use this review's findings to shape AI integration in nursing management and enhance patient support and interdisciplinary collaboration. Additionally, healthcare policymakers and AI developers can use this review to better understand and enhance AI-driven nursing practice in evolving healthcare systems.
Ethical approval
Because no human subjects were engaged in this study, no Institutional Review Board approval was required.
PROSPERO registration number
CRD42025649178
Funding
This work was funded by the
Author Statement
Neither this manuscript nor parts of it have been submitted elsewhere for publication.
CRediT authorship contribution statement
Jee Young Joo: Writing – review & editing, Writing – original draft, Validation, Supervision, Project administration, Methodology, Data curation, Conceptualization. Liu Megan F: Writing – review & editing, Writing – original draft, Validation, Supervision, Methodology, Formal analysis, Data curation, Conceptualization. Ho Benjamin: Writing – review & editing, Validation, Supervision, Project administration, Formal analysis, Conceptualization.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supporting information
Supplementary data associated with this article can be found in the online version at
Appendix A Supplementary material
Supplementary material
Table 1
| Authors
(year) Journal Setting(s) Country |
Study aim |
• Target Sample (sample size) • Technology Experience • Work Experience • Mean Age/Age Range • Sampling Method |
• Study Design • Qualitative research reporting framework • Theoretical framework • Ethical Considerations |
• Data Collection Methods • Data Analysis Method |
Major Nurses' Perceptions |
| Hassan & El-Ashry (2024)
BMC Nursing Four hospitals Egypt |
To investigate nurses' perception of AI's impact on professional identity, ethical considerations, and the shared meanings of trust, collaboration, and communication when working with AI systems |
• Intensive care unit nurses ( N = 10) • Yes • ≥ 2 years in critical care settings • 40.9 years • Purposive sampling |
• Qualitative phenomenological analysis • NR • Interpretive Phenomenological Analysis • Ethical approval of author's university & informed consent from nurses' participants |
• In-depth, semi-structured, one-on-one interviews • Thematic analysis |
• AI offers task automation but raises concerns about overreliance and the need for ongoing training. • Ethical considerations, clear communication, and collaboration were crucial. |
|
Helman et al. (2022)
International Journal of Medical Informatics University-affiliated hospital USA |
To elicit iterative design feedback from clinical end-users on an early GUI prototype display and inform the user-engaged design prototype utilizing previous experience with implementing new technologies into clinical workflow and user perspectives on GUI screen changes. |
• Nurses ( N = 18; 14 RNs, four NPs) • Yes • Median: 8 years • Median: 35 years • Purposive & Snowball Sampling |
• Qualitative focus group study • NR • NR • Approval of an IRB & informed consent obtained from nurses' participants |
• Virtual focus groups • Thematic analysis |
• Early engagement of nurses in AI design improves GUI usability • Early user-engaged design was useful in adjusting the GUI presentation of AI output |
|
King et al. (2023)
JAMIA Open University-affiliated hospital USA |
To explore how AI integration into handoff workflow influences situational awareness, assessment, monitoring, and communication goals of postoperative ward nurses |
• Nurses ( N = 11; seven PACU nurses & four ward nurses)• Yes • NR • NR • Convenience sampling |
• Qualitative study • COREQ checklist • NR • Approval of an IRB & informed consent obtained from nurses' participants |
• Semi-structured phone interviews & focus groups • Inductive-deductive reflexive thematic analysis |
• AI may address barriers to handoff effectiveness • AI may augment nursing decision-making and team handoff communication • EHR user experience and information overload are barriers to using AI during handoffs • AI tools improve postoperative handoff communication by identifying specific risks faced by patients |
|
Laukka et al. (2022)
Journal of Nursing Management University Hospital Finland |
To describe nurses' perceptions of AI's future role in specialized medical care |
• Nurses ( N = 20; 9 head nurses, 8 assistant head nurses, & 3 middle nurse managers)• Yes • 4.9–19.2 years of experience • 46–59 years • Purposive sampling |
• Qualitative study • COREQ checklist • NR • Approval from hospital ethics committee & informed consent obtained from nurses' participants |
• Semi-structured interviews & focus groups • Content analysis |
• AI will have a significant future role in specialized medical care, freeing up clinicians' time and enabling more patient interaction. • AI would handle routine documentation, allowing clinicians to focus on other demanding tasks. |
|
Rony et al. (2024a)
Heliyon Three tertiary-level hospitals Bangladesh |
To examine nurses' perceptions of AI's role in shaping the future of healthcare |
• RNs ( N = 23; 15 females, eight males)• Yes • 5–22 years of experience • 28–56 years old • Purposive sampling: minimum master's degree required |
• Qualitative study • COREQ checklist • NR • Approval from the ethics committee & informed consent were obtained |
• Semi-structured interview • Thematic analysis |
• AI integration is seen as an opportunity to improve patient care and streamline processes. • Concerns exist regarding AI's impact on nursing expertise and the unique human connection central to nursing. |
|
Rony et al. (2024b)
Health Science Reports One university hospital Bangladesh |
To gain insight into NP's perceptions and attitudes toward AI adoption in healthcare settings |
• RNs ( N = 37; 29 females, eight male) females • Yes • 3–30 years of experience • 28–50 years old • Purposive sampling(Maximum Variation and Expert Sampling) |
• Qualitative study with a descriptive and phenomenological approach • COREQ checklist • Technology Acceptance Model (TAM) • Approval from Institutional Review Committee at Bangladesh Open University & informed consent & confidentiality maintained |
• Semi-structured face-to-face or video interviews ( 50 −60 min)• Thematic analysis |
• Necessity of collaboration between policymakers and healthcare professionals is vital to balance innovation and patient safety. • Nurses were willing to adopt AI. • Concerns about integrating AI into hospital-established systems • Need for careful planning to avoid workflow disruptions |
|
Uymaz et al. (2024)
Environment and Social Psychology One research hospital Turkey |
To explore nurses' attitudes toward AI use for outpatients with chronic diseases |
• RNs (N = 5) • Yes • NR • 21–40 years old • Purposive sampling |
• Mixed method study with interview • NR • Unified Theory of Acceptance and the Use of Technology (UTAUT2) model • Approval from the research hospital |
• Individual interviews • NR |
• AI is considered suitable for nursing tasks involving managing the health of outpatients with chronic diseases. • Nurses have concerns about the reliability of ambulatory patient data. |
|
Wieben et al. (2024)
Journal of Nursing Scholarship University Hospital Midwest, USA |
To explore nurses' perspectives on Machine Learning Clinical Decision Support (ML CDS) design, development, implementation, and adoption |
• RNs ( N = 5) • Yes • Working experience range: NR • Age range: under 40 years (68.8 %) • 1–5 years (40 %), over 15 years (30 %) • Convenience sampling |
• Qualitative study • NR • Technology Acceptance Model • Approval from the Institutional Review Committee at Bangladesh Open University & informed consent & confidentiality were obtained. |
• Semi-structured individual interview using a secure video meeting platform (WebEx®) • Thematic analysis |
• The study shows that nurses were open to trying ML CDS if they understood how it could improve their practice and benefit their patients. • Involving nurses as active members of the design and development teams could enhance model design and adoption |
|
Zhai et al. (2022)
International Journal of Medical Informatics Tertiary hospital China |
To explore barriers and facilitators to the implementation of a CDSS from nurses' perspectives |
• Nurses (N = 21; nursing manager, bedside nurse, head nurse) • Yes • > 1 year of experience • NR • Purposive sampling |
• Qualitative study • FITT framework • Approved by the Ethics Review Committee and informed consent obtained from nurses' participants |
• In-depth interviews • Opening coding strategies |
• Barriers and facilitators to CDSS implementation include system, user, and organizational factors, which can largely fit into the FITT framework. • Interventions that foster cross-institutional collaboration rather than targeting single attributes would be key for successful implementation. |
Table 2
| Category | Main themes | Representative coding texts |
| Barriers | Ethical issues |
• Patients’ health data privacy and security issues ( Rony et al., 2024a; Rony et al., 2024b) • No clear set of ethical guidelines, standards, and political and legal regulations are provided ( Rony et al., 2024b; Zhai et al., 2022) • Reduces nurses’ critical thinking skills ( Hassan and El-Ashry, 2024; Helman et al., 2022; Laukka et al., 2022; Rony et al., 2024b; Wieben et al., 2024) • Health inequity: it may be possible for vulnerable populations to have unequal access to healthcare ( Rony et al., 2024b) |
| Increased workload |
• Additional education, time commitment, and new technical training regarding the use of AI tools are required ( Hassan and El-Ashry, 2024; Helman et al., 2022; Laukka et al., 2022; Uymaz et al., 2024; Zhai et al., 2022) • Continuous time-intensive and constant demand for the acquisition of skills for new AI tools and systems ( Laukka et al., 2022; Rony et al., 2024a; Rony et al., 2024b) • Frustration arising from balancing hands-on patient care with learning new AI tools and ensuring the accuracy of recorded patient data ( King et al., 2023; Uymaz et al., 2024; Wieben et al., 2024) | |
| Facilitators | Seamless & efficient patient care |
• AI tools may make it possible to predict potential deterioration or detect patient risks earlier based on patient data ( King et al., 2023; Rony et al., 2024a; Rony et al., 2024b; Wieben et al., 2024) • Guide clinical decision and supports nursing diagnosis ( Wieben et al., 2024) • Increase patient-centered care and patient safety. Reduces clinical errors ( Hassan and El-Ashry, 2024; Laukka et al., 2022; Wieben et al., 2024) • Ability to save time to concentrate and perform imperative tasks, thus ensuring enhanced patient care (Hassan & El- Laukka et al., 2022; Rony et al., 2024b) |
| Reinforcement rather than replacement |
• AI robots cannot replace human nurses; nurses have insights that cannot be detected by AI tools ( Laukka et al., (2022); Rony et al., 2024a; Wieben et al., (2024) • Support nurse-patient interaction and communication instead of replacing direct communication ( Laukka et al., 2022; Rony et al., 2024a; Rony et al., 2024b; Wieben et al., 2024) • Empathetic and holistic care is only possible by nurses; AI serves as an adjunct rather than a substitute ( Hassan and El-Ashry, 2024; Laukka et al., 2022; Rony et al., 2024a) | |
| AI as a future nursing care solution |
• AI tools are innovative and will make both nursing practices and patients’ care better ( Hassan and El-Ashry, 2024; King et al., 2023; Rony et al., 2024a; Rony et al., 2024b; Uymaz et al., 2024) • Nursing practices will be synergized with AI tools, thus enhancing the willingness to adopt AI tools ( Helman et al., 2022; Laukka et al., (2022) • Value of AI tools: supportive, reliable, and helpful to nurses ( Hassan and El-Ashry, 2024; King et al., 2023; Uymaz et al., 2024; Wieben et al., 2024; Zhai et al., 2022) |
© 2025 The Authors