Abstract
-
Purpose
This study examined the relationships among nurses’ readiness for artificial intelligence (AI), attitudes toward AI, and behavioral intention to use AI, focusing on clinical nurses in a tertiary hospital setting.
-
Methods
A cross-sectional descriptive study was conducted using an online self-report survey of 218 clinical nurses recruited through convenience sampling from a tertiary hospital in South Korea. AI readiness was measured using the Medical Artificial Intelligence Readiness Scale, attitudes toward AI were assessed using the Korean version of the General Attitudes toward Artificial Intelligence Scale, and behavioral intention was measured using items adapted from the Unified Theory of Acceptance and Use of Technology. Open-ended responses were summarized descriptively to explore expected AI applications.
-
Results
Clinical nurses demonstrated varying levels of AI readiness, attitudes toward AI, and behavioral intention to use AI, and these variables were positively correlated. Among AI readiness dimensions, ability and ethics tended to show stronger bivariate correlations with behavioral intention than vision. Hierarchical regression analysis indicated that attitudes toward AI were strongly associated with behavioral intention (β=.61, p<.001), whereas AI readiness factors showed weaker associations after attitudes were included. Open-ended responses suggested potential AI applications in both direct and indirect nursing care.
-
Conclusion
Attitudes toward AI were strongly associated with nurses’ behavioral intention to use AI. AI readiness dimensions, particularly ability and ethics, were also associated with behavioral intention in correlation analyses, underscoring the importance of practical competence and ethical awareness. These findings provide empirical evidence to inform AI-related education, clinical integration, and organizational support strategies in nursing.
-
Key Words: Artificial intelligence; Nurses; Health knowledge, attitudes, practice; Professional competence; Hospitals, teaching
INTRODUCTION
Artificial intelligence (AI) is increasingly being integrated into healthcare, including applications in diagnosis, prognosis prediction, image analysis, and personalized treatment planning [
1,
2]. In nursing practice, AI has demonstrated potential in areas such as patient monitoring, clinical decision support, and automated documentation [
3-
5], raising both expectations and concerns regarding changes in professional roles [
6] and ethical responsibilities [
7-
9]. Although international guidelines, including those from the International Council of Nurses, emphasize strengthening nurses’ digital competencies and AI-related education [
10,
11], as well as establishing ethical frameworks for AI [
12], the successful and safe implementation of AI ultimately depends on how clinical nurses perceive, prepare for, and engage with these technologies in real-world practice.
Research on AI in nursing has expanded rapidly in recent years; however, many studies have primarily focused on attitudes toward AI and intention to use AI in educational contexts involving nursing students [
13-
15]. Studies involving clinical nurses have gradually emerged and generally report positive perceptions of AI’s potential, alongside concerns regarding ethical issues and its impact on nursing practice [
3,
16,
17]. Additionally, education level and technological experience have been identified as factors associated with attitudes toward AI and intention to use AI [
14,
17,
18]. Among studies conducted in clinical settings, some have examined levels of awareness and attitudes [
19], and one study involving intensive care unit nurses explored the relationship between acceptance and readiness for AI [
20]. Nevertheless, these studies have often been limited to specific countries or clinical departments.
More recently, research has expanded to include topics such as attitudes and competencies, acceptance of organizational change, and experiences with generative AI across diverse populations, including clinical nurses [
21-
24]. Previous studies have reported associations between nurses’ attitudes toward AI and their self-efficacy and clinical reasoning competence [
21], as well as differences in readiness and acceptance according to job category and education level [
22,
24]. Experiences with generative AI have also been associated with more favorable attitudes toward AI and greater intention to use AI [
23]. However, despite this growing body of research, many studies have continued to rely on educational settings or community-based surveys [
13,
15,
21,
25-
27], which may not adequately reflect the experiences of nurses providing direct patient care in clinical environments.
Nurses, as frontline providers of patient care, are key agents in determining the success or failure of AI implementation and, as producers of clinical data and information records, influence AI performance and patient safety [
1,
2]. Therefore, clinical nurses’ AI readiness and attitudes should be regarded as essential factors for ensuring the safe and effective use of AI in practice, rather than merely reflecting levels of perception. Nevertheless, empirical studies that comprehensively examine readiness, attitudes, and intention to use AI remain limited.
From a theoretical perspective, this study is grounded in the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT), which conceptualize attitudes as a proximal determinant of behavioral intention toward technology use [
28,
29]. Within this framework, readiness-related factors—including perceived ability, ethical awareness, and contextual understanding of AI—are assumed to be conceptually related to nurses’ evaluative attitudes toward AI. Attitudes toward AI are, in turn, associated with behavioral intention. Accordingly, attitudes may function as a key conceptual link between AI readiness and behavioral intention in clinical nursing practice. This framework guided the simultaneous examination of AI readiness, attitudes toward AI, and behavioral intention in the present study.
Therefore, this study examined AI readiness, attitudes toward AI, and behavioral intention to use AI among clinical nurses in a tertiary hospital in South Korea. Specifically, the study aimed to (1) assess levels of AI readiness and attitudes toward AI, (2) examine correlations among AI readiness, attitudes, and behavioral intention, and (3) identify factors associated with behavioral intention using hierarchical regression analysis. In addition, open-ended responses were analyzed to identify expected areas of AI utilization in clinical nursing practice that may not be fully captured through quantitative measures.
METHODS
1. Study Design
This was a cross-sectional descriptive correlational study. The study is reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines.
2. Setting and Samples
Participants were clinical nurses with at least 1 year of work experience at a tertiary hospital in South Korea. Eligible participants were licensed nurses currently working in clinical departments who had provided direct patient care for at least 1 year. Nurses who were on leave or had resigned, those not engaged in clinical duties, those working exclusively in research or education roles, and those with less than 1 year of total clinical experience were excluded. These criteria were established to ensure that participants could reliably report perceptions of and attitudes toward AI in the current clinical environment.
The sample size was calculated using G*Power 3.1.9.7 based on the primary analysis of hierarchical multiple regression. Assuming a significance level of .05, power of .80, a medium effect size (f²=0.15), and up to 10 predictors, the minimum required sample size was 118. In this calculation, predictors were defined as conceptually distinct variables at the design stage, and categorical variables were treated as single predictors for the purpose of sample size estimation. To account for potential dropout and to ensure sufficient power for additional subgroup and comparative analyses, 220 participants were recruited. After excluding two responses that did not meet the inclusion criteria, data from 218 participants were included in the final analysis.
3. Measurements
1) Demographic characteristics
General characteristics included gender, age, education level, graduate education experience, clinical department, job position, and years of clinical experience.
2) AI readiness
AI readiness was measured using the Medical Artificial Intelligence Readiness Scale for Medical Students (MAIRS-MS), developed by Karaca et al. [
25], which was translated with permission from the original authors and underwent content validity assessment. MAIRS-MS-derived readiness constructs have also been examined among healthcare professionals in clinical settings [
22], suggesting potential applicability beyond student populations. Content validity was evaluated by an expert panel consisting of one nursing faculty member and clinical nurses with more than 10 years of clinical experience who held at least a master’s degree in nursing and had prior experience related to AI use or implementation in hospital settings. The translated version demonstrated good content validity, with a scale-level content validity index average of 0.94. The instrument consists of 22 items across four dimensions—cognition, ability, vision, and ethics—rated on a 5-point Likert scale (1=strongly disagree, 5=strongly agree), with higher scores indicating greater readiness to use AI in healthcare. Cronbach’s α was .87 in the original study and .94 in this study.
3) Attitudes toward AI
Attitudes toward AI were measured using the Korean version of the General Attitudes toward Artificial Intelligence Scale (GAAIS-K) [
13], translated from the original GAAIS [
30]. GAAIS-K consists of 18 items, including 11 positive items and 7 negative items, rated on a 5-point Likert scale. Negative items were reverse-coded according to the scoring guidelines so that higher scores indicated more favorable attitudes on both subscales. For analysis, attitudes toward AI were treated as a single composite score derived from all 18 items, and positive and negative subscale scores were also calculated separately. This composite score was used to capture the overall tendency of attitudes toward AI across items. Cronbach’s α was .88 for the positive subscale and .83 for the negative subscale in the original scale, .86 and .74 in GAAIS-K, and .83 and .78 in this study.
4) Behavioral intention to use AI
Behavioral intention to use AI was measured using three items [
31] adapted for the healthcare context based on the UTAUT [
29]. Items were rated on a 5-point Likert scale, with higher scores indicating stronger behavioral intention to use AI. Cronbach’s α was .90 in the original scale, .69 in the adapted version, .72 in a previous study [
32], and .71 in this study.
5) Open-ended survey items
In addition to the standardized instruments, participants were asked whether they had experience using AI-based programs or systems in clinical practice. Open-ended questions were included to explore clinical AI experiences and expectations, including experiences with AI use in clinical settings and nursing tasks for which AI was expected to be useful. Responses were grouped into categories based on similarity of meaning, and frequencies were calculated.
4. Data Collection
Data were collected from November 18 to December 11, 2024. Participants were recruited through convenience sampling using a recruitment notice posted on the hospital intranet bulletin board, which was accessible to all nurses employed at the study institution. Nurses who met the inclusion criteria and voluntarily agreed to participate completed the online survey.
The inclusion criteria were as follows: (1) registered nurses currently working at a tertiary hospital in South Korea, (2) nurses who had provided direct patient care in clinical settings for at least 1 year, and (3) individuals who understood the study purpose and provided informed consent. The exclusion criteria were as follows: (1) nurses not currently engaged in direct clinical practice, (2) nurses with less than 1 year of clinical experience, and (3) incomplete survey responses or withdrawal of consent. Participants received detailed information about the study purpose and procedures, provided electronic informed consent, and completed a structured self-report online questionnaire. The survey required approximately 10 minutes to complete. Measurement instruments were used after permission had been obtained from the original authors and translators.
5. Ethical Considerations
This study was approved by the Institutional Review Board of Seoul National University Hospital (No. H-2410-122-1579). All participants were informed of the study purpose and procedures and voluntarily agreed to participate through an online consent process.
6. Data Analysis
Data were analyzed using IBM SPSS ver. 25.0 (IBM Corp., Armonk, NY, USA) and R version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria). To examine the structural validity of the MAIRS-MS, confirmatory factor analysis was conducted using robust maximum likelihood estimation to account for potential non-normality. Differences in the major variables according to general characteristics were analyzed using an independent t-test and one-way analysis of variance. When the assumption of homogeneity of variance was violated, Welch’s test was applied. Pearson correlation coefficients were calculated to examine relationships among the key variables.
Hierarchical regression diagnostic checks were performed to assess residual independence (Durbin-Watson test), multicollinearity (variance inflation factor [VIF]), linearity, and homoscedasticity. Hierarchical linear regression analysis was conducted using the mean score of the three behavioral intention to use AI items as the dependent variable. General characteristics, AI readiness, and attitudes toward AI were entered sequentially as predictors. Categorical variables were dummy-coded (0=reference group, 1=comparison group), with the most frequent category used as the reference group. Based on TAM and UTAUT, which conceptualize attitudes as a proximal determinant of behavioral intention, demographic variables were entered as control variables in the first block, followed by AI readiness as a more distal factor and attitudes toward AI in the final block [
28,
29]. Open-ended responses were summarized descriptively by reporting categories and frequencies to provide contextual information. Responses were grouped into categories based on similarity of meaning. Categorization was refined through repeated review and discussion between the researcher and a nursing faculty member until consensus was reached. These data were used to complement the quantitative findings and were not subjected to in-depth qualitative analysis.
RESULTS
1. General Characteristics of the Participants
A total of 218 nurses participated in the study, and most were women (n=207, 95.0%). The largest age group was nurses in their 30s (n=85, 39.0%), followed by those in their 20s (n=77, 35.3%) and those aged 40 years or older (n=56, 25.7%); the mean age was 34.35 years (standard deviation [SD], 7.31 years). Most participants held a bachelor’s degree (82.1%), whereas 17.9% held a master’s degree or higher. Among the participants, 31.2% had graduate-level education experience.
The largest proportion of participants worked in wards (52.3%), followed by intensive care units (20.6%), outpatient clinics (11.5%), emergency rooms (4.6%), operating rooms (4.6%), and other departments (6.4%). Most participants were staff nurses (96.8%), whereas 3.2% were head nurses. The mean duration of clinical experience was 10.99 years (SD, 7.43 years), and 52.3% had less than 10 years of experience. Nurses with AI experience accounted for 10.1% of the sample (
Table 1).
The mean AI readiness score was 67.56 (SD, 12.97), and subscale scores for cognition, ability, vision, and ethics are presented in
Table 2. The mean scores for attitudes toward AI and behavioral intention to use AI were 66.16 (SD, 7.57) and 11.93 (SD, 1.94), respectively.
To examine the structural validity of the MAIRS-MS in this sample, confirmatory factor analysis was additionally conducted. The analysis generally supported the original four-factor structure, with fit indices in a marginal but acceptable range (χ²/df=2.54, comparative fit index=.898, Tucker-Lewis index=.884, root mean square error of approximation=.084, and standardized root mean square residual=.067). All standardized factor loadings were statistically significant and ranged from 0.53 to 0.91 across dimensions. Composite reliability values for the four factors ranged from 0.86 to 0.92, exceeding the threshold of .70. Average variance extracted values ranged from 0.49 to 0.68, which were at or near the recommended threshold of .50, supporting convergent validity.
3. Differences in Participants’ AI Readiness, Attitudes toward AI, and Behavioral Intention to Use AI according to General Characteristics
Differences in AI readiness, attitudes toward AI, and behavioral intention to use AI according to general characteristics are presented in
Table 1. Significant differences according to education level were observed for both AI readiness (F=3.84,
p=.023) and attitudes toward AI (F=5.89,
p=.003). Nurses with graduate education experience had significantly higher scores for AI readiness (t=3.67,
p<.001), attitudes toward AI (t=2.87,
p=.005), and behavioral intention to use AI (t=2.10,
p=.037).
Correlations among the major variables are presented in
Table 3. AI readiness was significantly and positively correlated with behavioral intention to use AI (r=.31,
p<.001). The strongest correlation with behavioral intention was observed for attitudes toward AI (r=.64,
p<.001). AI readiness was also significantly and positively correlated with attitudes toward AI (r=.30,
p<.001).
At the subscale level, among the AI readiness dimensions, ability showed the strongest correlation with behavioral intention (r=.36, p<.001), followed by ethics (r=.30, p<.001), vision (r=.19, p<.01), and cognition (r=.18, p<.01). For attitudes toward AI, the positive attitude subscale (r=.66, p<.001) exhibited a stronger positive correlation with behavioral intention than the negative attitude subscale (r=.40, p<.001). This positive correlation reflects the reverse coding of the negative attitude items, such that higher scores indicate more favorable attitudes toward AI.
5. Predictors of Behavioral Intention to Use AI
Hierarchical regression analysis was conducted to identify factors associated with behavioral intention to use AI (
Table 4). This analytic approach was theoretically grounded in the TAM and the UTAUT [
28,
29]. By entering readiness variables first and attitudes subsequently, the analysis examined whether attitudes provided incremental explanatory power for behavioral intention, consistent with the conceptual pathway suggested in previous literature. This analytic strategy was intended to assess the sequential contributions of readiness and attitudes after controlling for general characteristics, rather than to formally test mediation effects.
In model 1, general characteristics were entered as control variables, and the explanatory power of the model was low (adjusted R²=.01). Categorical variables were dummy-coded, with the most frequent category used as the reference group. None of the general characteristics, including age, showed a significant association with behavioral intention in this model. When AI readiness was added in model 2, the explanatory power increased significantly (adjusted R²=.09, ΔR²=.08, p<.001), indicating that readiness contributed additional explanatory value beyond general characteristics. In model 3, attitudes toward AI were additionally entered, resulting in the highest explanatory power (adjusted R²=.43, ΔR²=.32, p<.001). Attitudes toward AI emerged as a factor strongly associated with behavioral intention to use AI (β=.61, p<.001). Although the effect of AI readiness was attenuated after the inclusion of attitudes, it remained statistically significant (B=0.17, β=.16, p=.005). Gender was not significant in models 1 and 2 but emerged as a significant predictor in the final model.
Regression diagnostics indicated no violation of model assumptions. The Durbin-Watson statistic was 2.03, and VIF values ranged from 1.05 to 7.44. Although the VIF values for age (7.44) and the clinical experience group of 20 years or more (7.10) were higher than those of the other predictors, all VIF values remained below the commonly accepted threshold of 10, suggesting that severe multicollinearity was unlikely. Despite the elevated VIF values for these two predictors, they were retained for theoretical reasons, and the coefficient estimates remained stable. Residual diagnostics confirmed that the assumptions of normality (Shapiro-Wilk p=.54) and homoscedasticity (Breusch-Pagan p=.59) were satisfied. The maximum Cook’s distance was 0.18, indicating that no influential observations substantially affected model stability.
6. Expected Areas of AI Use in Clinical Nursing Practice
Table 5 presents the descriptive results of the open-ended survey questions regarding nurses’ experiences with AI and their expectations for AI applications in clinical nursing practice. Nurses identified potential AI applications in both direct patient care and indirect nursing activities.
In addition to the quantitative findings, open-ended responses indicated that nurses emphasized the need for adequate education, institutional support, and safeguards to ensure patient safety during AI implementation, while some also expressed positive expectations regarding improved work efficiency.
DISCUSSION
This study has both academic and practical significance as one of the few studies to examine the relationships among AI readiness, attitudes toward AI, and behavioral intention to use AI among clinical nurses in a tertiary hospital. Previous studies primarily targeted nursing students [
13-
15], and some focused on primary healthcare workers [
26], thereby reflecting educational expectations or community healthcare contexts. Studies involving clinical nurses have recently emerged; however, most have reported only levels of awareness and attitudes [
19], whereas one study involving intensive care unit nurses examined AI readiness and acceptance but was limited to a specific unit [
20]. In contrast, the present study is distinguished by its comprehensive assessment of AI readiness, attitudes toward AI, and behavioral intention to use AI among nurses working across diverse departments in a tertiary hospital.
In this study, clinical nurses demonstrated an above-midpoint level of AI readiness, indicating a generally favorable level of perceived readiness for AI-related practice. This level of readiness should be interpreted in light of the characteristics of the measurement instrument, which conceptualizes AI readiness as a multidimensional construct encompassing perceived ability, ethical considerations, vision, and acceptance of AI, rather than as direct evidence of extensive hands-on experience. Accordingly, the observed readiness reflects nurses’ evaluative and perceptual preparedness toward AI within the clinical context. Consistent with the relatively higher scores observed in the ethics dimension, the open-ended responses revealed positive expectations regarding improved work efficiency and clinical support for patient monitoring and risk prediction, all of which are closely related to patient safety in clinical care. These findings suggest that nurses’ AI readiness reflects contextual and perception-based preparedness that coexists with careful consideration of clinical responsibility and safety, in line with the Code of Ethics for Nurses [
33].
A previous study using the same instrument reported higher AI readiness among physicians than among nurses [
22]. This difference may partly reflect greater opportunities for physicians to engage directly with AI systems in diagnostic and decision-making processes. In contrast, AI-related nursing research has more often focused on attitudes and adaptation rather than readiness itself [
13-
21]. Given potential differences in professional roles and practice environments, it is important to consider whether the MAIRS-MS adequately captures AI readiness in clinical nurses before interpreting these comparisons.
Although the MAIRS-MS was originally developed for medical students, the present findings indicate that its four-factor structure was generally retained in a clinical nursing population. The overall model fit indices were slightly below conventional thresholds, which may reflect contextual differences between student and practicing clinician populations. Nevertheless, the factor loadings and convergent validity indicators supported the structural adequacy of the instrument. These findings are consistent with recent studies applying MAIRS-related constructs to healthcare professionals beyond student samples [
22]. Together, these results suggest that the MAIRS-MS may be cautiously applied to clinical nurses, although further validation in diverse professional samples is warranted.
Among the AI readiness dimensions, ability and ethics showed relatively stronger associations with behavioral intention than vision in the correlation analysis. This finding suggests that clinical nurses place greater importance on the practical feasibility of AI use and responsibility for patient safety than on future-oriented expectations. The vision dimension showed a relatively weaker association than the other dimensions, indicating that long-term or abstract perceptions of AI may not yet have translated into concrete intentions in clinical practice. This pattern contrasts with findings in medical students, for whom future-oriented perceptions played a more prominent role [
25], and with findings in intensive care unit nurses, in whom perceived usefulness and facilitating conditions were emphasized [
20]. Overall, these results suggest that AI acceptance among clinical nurses is grounded primarily in immediate applicability and ethical considerations rather than in visionary expectations, reflecting the realities of frontline nursing practice.
Graduate education experience was associated with higher AI readiness and more positive attitudes toward AI; however, it did not remain significant in the final regression model. This pattern is consistent with previous studies suggesting that educational exposure may influence behavioral intention indirectly through attitudes or perceived competence rather than as an independent determinant [
14]. In this context, graduate education may contribute to behavioral intention by shaping nurses’ cognitive and evaluative frameworks regarding AI, including conceptual understanding, critical appraisal skills, and ethical awareness. These educational influences are more likely to be reflected in attitudes toward AI, which are more proximally associated with behavioral intention. Although this pattern suggests a potential indirect pathway, it should be interpreted cautiously because formal mediation analysis was not conducted. These findings imply that formal education alone may be insufficient without corresponding changes in attitudes and readiness, highlighting the need for additional institutional and educational support to strengthen nurses’ AI readiness in clinical settings.
Most demographic variables, such as age, department, and job position, were not significantly associated with behavioral intention. Notably, gender was not significant in Models 1 and 2 but emerged as a significant predictor in the final model after accounting for psychological factors such as attitudes. This finding suggests that the association between gender and behavioral intention to use AI may depend on the inclusion of psychological variables and should therefore be interpreted as an adjusted association within the full model rather than as a direct effect. Overall, these results are consistent with previous studies reporting inconsistent or negligible effects of demographic factors on AI acceptance, suggesting that individual perceptions and attitudes play a more central role than sociodemographic characteristics [
16,
17]. However, because male participants represented only a small proportion of the sample (n=11, 5.0%), this finding should be interpreted cautiously, as the substantial gender imbalance may limit the stability and generalizability of this association.
In the hierarchical regression analysis, attitudes toward AI showed a strong association with behavioral intention to use AI in the final model (β=.61,
p<.001). This finding is consistent with the TAM and the UTAUT, which conceptualize attitudes as a proximal factor associated with behavioral intention [
28,
29]. In line with this theoretical framework, previous empirical studies have also identified attitudes toward AI as an important factor associated with behavioral intention to use AI [
13-
16].
However, the emphasis placed on AI acceptance appears to vary across participant groups. Previous studies have shown that vision and innovation acceptance were more salient among medical students [
25], whereas educational experience and the mediating role of attitudes were emphasized among nursing students [
13,
14]. Community nurses have also highlighted the perceived impact of AI on professional practice [
26]. In contrast, the present findings showed that the ability and ethics dimensions were more strongly associated with behavioral intention than the vision dimension among nurses in a tertiary hospital. This pattern suggests that AI acceptance in this group may be more closely related to immediate applicability and patient safety considerations. However, because these dimensions were examined using correlation analyses and were not simultaneously included in the regression model, these findings should be interpreted as reflecting differences in association strength rather than differential causal influence. These results underscore the context-dependent nature of AI acceptance and highlight the clinical realities of frontline nursing practice.
This pragmatic orientation is further reflected in the specific areas of expected AI utilization identified in the open-ended responses. Nurses identified potential AI applications in both direct and indirect nursing activities. Specifically, the prevention of medication errors and support for patient monitoring correspond to direct nursing activities related to patient safety, whereas automated documentation and nurse staffing support represent indirect activities that enhance efficiency and administrative management. Previous literature has similarly emphasized clinical applications such as monitoring, documentation automation, and prediction of nursing outcomes [
34], and exploratory studies have reported high perceived usefulness of AI for repetitive and time-consuming tasks [
35]. The present findings are consistent with this evidence, indicating that nurses view AI as augmenting their work and enhancing patient safety rather than replacing nursing roles.
Overall, this study identified relationships among AI readiness, attitudes toward AI, and behavioral intention to use AI among clinical nurses and additionally explored expected areas of AI utilization in clinical nursing practice through open-ended responses. Unlike previous studies focusing on students or specific departments, this study provides empirical evidence from nurses directly involved in patient care across diverse clinical settings. These findings provide foundational data for developing strategies to strengthen nurses’ AI competency through educational and institutional support.
This study has several limitations. First, the single-center, cross-sectional design limits the generalizability of the findings, and the results may not be directly applicable to nurses working in healthcare settings with less access to digital infrastructure. Because participation was voluntary and the survey was conducted online, nurses with greater interest in AI may have been more likely to respond, potentially leading to overestimation of AI readiness. However, only 10.1% of participants reported direct clinical experience with AI, suggesting that the sample largely consisted of nurses with limited direct exposure to AI in practice. In addition, perceptions of AI are evolving rapidly; therefore, these findings reflect a specific time point and may change over time. Further research should adopt longitudinal and multicenter designs.
Although the sample size met conventional minimum criteria for confirmatory factor analysis, it may still be considered modest for achieving stable structural validation. Therefore, the confirmatory factor analysis findings should be interpreted cautiously, and further validation in larger and more diverse nursing populations is warranted. In addition, the effect of AI readiness on behavioral intention was attenuated after attitudes toward AI were entered into the hierarchical regression model, suggesting a potential indirect role of attitudes. However, formal mediation analysis was not conducted. Accordingly, this finding should be interpreted cautiously and regarded as exploratory. Future studies may further examine this potential mechanism using mediation or path-analytic approaches, such as structural equation modeling.
Finally, the analysis of the open-ended responses was limited to basic categorization. Future studies using in-depth interviews or focus groups are warranted to provide a deeper understanding of nurses’ perceptions of AI.
CONCLUSION
The findings of this study indicate that attitudes toward AI were strongly associated with behavioral intention to use AI among clinical nurses in a tertiary hospital. AI readiness and graduate education experience were related to attitudes toward AI, suggesting that these factors may contribute to behavioral intention indirectly rather than exerting a direct effect. In the correlation analyses, the ability and ethics dimensions of AI readiness tended to show stronger associations with behavioral intention in clinical contexts, suggesting that nurses’ perceptions may be grounded primarily in practical needs and safety considerations. In the open-ended responses, nurses expressed expectations that AI could contribute to both direct and indirect nursing activities, particularly in relation to patient safety and work efficiency. Taken together, these findings suggest that educational and organizational strategies aligned with clinical nurses’ practical needs and ethical considerations may facilitate the integration of AI into nursing practice, based on the associations observed in this study.
-
CONFLICTS OF INTEREST
The authors declared no conflict of interest.
-
AUTHORSHIP
Study conception and design acquisition - JES and YHP; analysis - JES; interpretation of the data - JES; and drafting or critical revision of the manuscript for important intellectual content - JES and YHP.
-
FUNDING
This work was supported by the 2024 Graduate Student Research Grant from the Research Institute of Nursing Science, Seoul National University.
-
ACKNOWLEDGEMENT
ChatGPT (OpenAI, San Francisco, CA, USA) was used to refine the clarity and fluency of the English expressions during manuscript preparation.
-
DATA AVAILABILITY STATEMENT
The data can be obtained from the corresponding author.
Table 1.General Characteristics of Participants and Differences in AI Readiness, Attitudes toward AI, and Behavioral Intention to Use AI (N=218)
|
Characteristics |
Categories |
n (%) |
M±SD |
AI readiness |
Attitudes toward AI |
Behavioral intention to use AI |
|
|
M±SD |
t or F (p) |
M±SD |
t or F (p) |
M±SD |
t or F (p) |
|
|
Sociodemographic characteristics |
|
|
|
|
|
|
|
|
|
|
|
Gender |
Women |
207 (95.0) |
- |
67.26±12.74 |
–1.18 (.265) |
65.91±7.43 |
–1.76 (.107) |
11.93±1.98 |
0.32 (.751) |
|
|
Men |
11 (5.0) |
|
73.18±16.44 |
|
70.73±8.92 |
|
11.82±1.08 |
|
|
|
Age |
20s |
77 (35.3) |
34.35±7.31 |
68.74±11.20 |
0.50 (.605) |
65.66±6.78 |
0.68 (.506) |
11.90±1.90 |
1.13 (.325) |
|
|
30s |
85 (39.0) |
|
67.06±12.54 |
|
66.91±8.07 |
|
12.14±2.00 |
|
|
|
≥40 |
56 (25.7) |
|
66.70±15.72 |
|
65.70±7.83 |
|
11.64±1.89 |
|
|
|
Education |
Bachelor’s |
179 (82.1) |
- |
66.58±12.83 |
3.84 (.023) |
65.60±6.99 |
5.89 (.003) |
11.84±1.85 |
1.55 (.215) |
|
|
Master’s |
37 (17.0) |
|
71.41±12.78 |
|
69.38±9.24 |
|
12.41±2.33 |
|
|
|
Doctoral |
2 (0.9) |
|
84.00±7.07 |
|
56.00±1.41 |
|
11.00±1.41 |
|
|
|
Graduate education experience |
Yes |
68 (31.2) |
- |
72.09±11.99 |
3.67 (<.001) |
68.46±8.40 |
2.87 (.005) |
12.35±2.09 |
2.10 (.037) |
|
|
No |
150 (68.8) |
|
65.51±12.91 |
|
65.11±6.94 |
|
11.73±1.85 |
|
|
|
Job-related characteristics |
|
|
|
|
|
|
|
|
|
|
|
Department |
Wards |
114 (52.3) |
- |
66.71±14.35 |
0.93 (.395) |
65.30±7.70 |
1.82 (.164) |
11.85±1.91 |
0.48 (.621) |
|
|
ICUs |
45 (20.6) |
|
67.16±11.12 |
|
66.47±7.42 |
|
12.18±2.01 |
|
|
|
Others |
59 (27.1) |
|
69.51±11.34 |
|
67.58±7.31 |
|
11.88±1.97 |
|
|
|
Job position |
Staff nurse |
211 (96.8) |
- |
67.74±12.84 |
0.91 (.395) |
66.13±7.60 |
–0.27 (.796) |
11.91±1.94 |
–0.44 (.676) |
|
|
Head nurse |
7 (3.2) |
|
62.00±16.47 |
|
66.86±6.99 |
|
12.29±2.21 |
|
|
|
Years of clinical experience |
<10 |
114 (52.3) |
10.99±7.43 |
68.96±11.60 |
1.97 (.142) |
66.45±7.09 |
0.30 (.738) |
12.00±1.97 |
1.24 (.291) |
|
|
10–19 |
73 (33.5) |
|
65.15±14.72 |
|
66.08±8.09 |
|
12.03±1.89 |
|
|
|
≥20 |
31 (14.2) |
|
68.06±12.97 |
|
65.26±8.12 |
|
11.42±1.95 |
|
|
|
AI-related characteristics |
|
|
|
|
|
|
|
|
|
|
|
AI experience |
Yes |
22 (10.1) |
- |
68.41±16.50 |
–0.26 (.797) |
67.14±9.32 |
–0.53 (.601) |
12.05±2.13 |
–0.28 (.783) |
|
|
No |
196 (89.9) |
|
67.46±12.56 |
|
66.05±7.36 |
|
11.91±1.92 |
|
|
Table 2.Descriptive Statistics for AI Readiness, Attitudes toward AI, and Behavioral Intention to Use AI (N=218)
|
Variables |
No. of items |
Range |
M±SD |
|
AI readiness |
22 |
24–109 |
67.56±12.97 |
|
Cognition |
8 |
9–39 |
21.50±5.19 |
|
Ability |
8 |
8–40 |
26.09±5.66 |
|
Vision |
3 |
3–15 |
9.15±2.41 |
|
Ethics |
3 |
3–15 |
10.82±2.16 |
|
Attitudes toward AI |
18 |
39–86 |
66.16±7.57 |
|
Positive |
11 |
26–55 |
43.72±4.99 |
|
Negative |
7 |
11–35 |
22.44±3.87 |
|
Behavioral intention to use AI |
3 |
6–15 |
11.93±1.94 |
Table 3.Correlations among AI Readiness, Attitudes toward AI, and Behavioral Intention to Use AI (N=218)
|
Variables |
AI readiness |
Attitudes toward AI |
Behavioral intention to use AI |
|
r (p) |
|
AI readiness |
1 |
.30 (<.001) |
.31 (<.001) |
|
Cognition |
|
.19 (.004) |
.18 (.009) |
|
Ability |
|
.35 (<.001) |
.36 (<.001) |
|
Vision |
|
.12 (.090) |
.19 (.005) |
|
Ethics |
|
.27 (<.001) |
.30 (<.001) |
|
Attitudes toward AI |
|
1 |
.64 (<.001) |
|
Positive |
|
|
.66 (<.001) |
|
Negative |
|
|
.40 (<.001) |
|
Behavioral intention to use AI |
|
|
1 |
Table 4.Hierarchical Regression Analysis Predicting Behavioral Intention to Use AI (N=218)
|
Variables |
Step 1 |
Step 2 |
Step 3 |
|
B |
SE |
β |
p
|
B |
SE |
β |
p
|
B |
SE |
β |
p
|
|
General characteristics |
|
|
|
|
|
|
|
|
|
|
|
|
|
Gender |
–0.13 |
0.20 |
–.04 |
.519 |
–0.21 |
0.20 |
–.07 |
.279 |
–0.38 |
0.16 |
–.13 |
.017 |
|
Age |
–0.01 |
0.02 |
–.17 |
.364 |
–0.01 |
0.02 |
–.09 |
.630 |
–0.00 |
0.01 |
–.04 |
.755 |
|
Graduate education experience |
0.26 |
0.10 |
.19 |
.009 |
0.15 |
0.10 |
.10 |
.147 |
0.01 |
0.08 |
.01 |
.901 |
|
Department |
|
|
|
|
|
|
|
|
|
|
|
|
|
ICUs |
0.08 |
0.12 |
.05 |
.489 |
0.09 |
0.11 |
.06 |
.399 |
0.05 |
0.09 |
.03 |
.543 |
|
Others |
0.07 |
0.11 |
.05 |
.533 |
0.02 |
0.10 |
.01 |
.847 |
–0.10 |
0.08 |
–.07 |
.221 |
|
Job position |
0.08 |
0.26 |
.02 |
.771 |
0.20 |
0.25 |
.05 |
.432 |
0.16 |
0.20 |
.04 |
.436 |
|
Years of clinical experience |
|
|
|
|
|
|
|
|
|
|
|
|
|
10–19 |
0.12 |
0.17 |
.08 |
.508 |
0.12 |
0.17 |
.09 |
.458 |
0.12 |
0.13 |
.08 |
.389 |
|
≥20 |
–0.03 |
0.33 |
–.02 |
.923 |
–0.10 |
0.32 |
–.05 |
.760 |
–0.02 |
0.25 |
–.01 |
.927 |
|
AI experience |
0.05 |
0.15 |
.02 |
.760 |
0.03 |
0.14 |
.01 |
.831 |
–0.02 |
0.11 |
–.01 |
.864 |
|
AI readiness |
|
|
|
|
0.33 |
0.07 |
.31 |
<.001 |
0.17 |
0.06 |
.16 |
.005 |
|
Attitudes toward AI |
|
|
|
|
|
|
|
|
0.94 |
0.09 |
.61 |
<.001 |
|
Model fit |
Adj.R²=.01 |
Adj.R²=.09; ΔR²=.08 |
Adj.R²=.43; ΔR²=.32 |
Table 5.Descriptive Results of Open-Ended Survey Questions on AI Use in Nursing Practice
|
Categories |
n (%) |
Example quotes |
|
AI usage experiences in clinical settings (n=21)†
|
|
|
|
Patient risk/prognosis prediction |
7 (33.3) |
“Patient risk analysis,” |
|
“Deterioration prediction” |
|
Work support |
6 (28.6) |
“Diagnostic prediction,” |
|
“Triage assistance” |
|
Imaging interpretation support |
5 (23.8) |
“Chest X-ray AI reading,” |
|
“Automated measurement of lung fields” |
|
Communication/information retrieval |
3 (14.3) |
“English translation and clinical knowledge lookup,” |
|
“Evidence summarization” |
|
Nursing tasks in which AI is expected to be useful (n=218) |
|
|
|
Patient monitoring/prediction |
61 (28.0) |
“Prediction of falls and pressure ulcers,” |
|
“Early detection of patient deterioration” |
|
Test result/imaging analysis |
49 (22.5) |
“Interpretation of test results,” |
|
“Assistance with imaging tests” |
|
Staffing/work allocation |
37 (17.0) |
“Efficient allocation of nursing staff,” |
|
“Reallocation according to patient condition” |
|
Medication/administration |
28 (12.8) |
“Drug interaction check,” |
|
“Adjustment of medication timing” |
|
Nursing records/documentation |
21 (9.6) |
“Automated nursing records,” |
|
“Auto-entry in charts” |
|
Unknown/no response |
18 (8.3) |
“Not sure yet,” |
|
“Nothing comes to mind” |
|
Others |
4 (1.8) |
“Support for medical decision-making,” |
|
“Searching medical records,” |
|
“Health consultation chatbot” |
REFERENCES
- 1. Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44-56. https://doi.org/10.1038/s41591-018-0300-7
- 2. Davenport T, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthc J. 2019;6(2):94-8. https://doi.org/10.7861/futurehosp.6-2-94
- 3. Ronquillo CE, Peltonen LM, Pruinelli L, Chu CH, Bakken S, Beduschi A, et al. Artificial intelligence in nursing: priorities and opportunities from an international invitational think-tank of the Nursing and Artificial Intelligence Leadership Collaborative. J Adv Nurs. 2021;77(9):3707-17. https://doi.org/10.1111/jan.14855
- 4. Robert N. How artificial intelligence is changing nursing. Nurs Manage. 2019;50(9):30-9. https://doi.org/10.1097/01.NUMA.0000578988.56622.21
- 5. von Gerich H, Moen H, Block LJ, Chu CH, DeForest H, Hobensack M, et al. Artificial Intelligence-based technologies in nursing: a scoping literature review of the evidence. Int J Nurs Stud. 2022;127:104153. https://doi.org/10.1016/j.ijnurstu.2021.104153
- 6. Blease C, Kaptchuk TJ, Bernstein MH, Mandl KD, Halamka JD, DesRoches CM. Artificial intelligence and the future of primary care: exploratory qualitative study of UK general practitioners’ views. J Med Internet Res. 2019;21(3):e12802. https://doi.org/10.2196/12802
- 7. Lysaght T, Lim HY, Xafis V, Ngiam KY. AI-assisted decision-making in healthcare. Asian Bioeth Rev. 2019;11(3):299-314. https://doi.org/10.1007/s41649-019-00096-0
- 8. Chen IY, Pierson E, Rose S, Joshi S, Ferryman K, Ghassemi M. Ethical machine learning in healthcare. Annu Rev Biomed Data Sci. 2021;4:123-44. https://doi.org/10.1146/annurev-biodatasci-092820-114757
- 9. Wysocki O, Davies JK, Vigo M, Armstrong AC, Landers D, Lee R, et al. Assessing the communication gap between AI models and healthcare professionals: explainability, utility and trust in AI-driven clinical decision-making. Artif Intell. 2023;316:103839. https://doi.org/10.1016/j.artint.2022.103839
- 10. International Council of Nurses. Digital health transformation and nursing practice. Geneva: International Council of Nurses; 2023.
- 11. Lattuca M, Maratta D, Beffert U, Chevrier A, Winer L. Healthcare AI: a revised Quebec framework for nursing education. Qual Adv Nurs Educ. 2023;9(3):Article 2. https://doi.org/10.17483/2368-6669.1408
- 12. Jobin A, Ienca M, Vayena E. The global landscape of AI ethics guidelines. Nat Mach Intell. 2019;1(9):389-99. https://doi.org/10.1038/s42256-019-0088-2
- 13. Seo YH, Ahn J. The validity and reliability of the Korean version of the General Attitudes towards Artificial Intelligence Scale for nursing students. J Korean Acad Soc Nurs Educ. 2022;28(4):357-67. https://doi.org/10.5977/jkasne.2022.28.4.357
- 14. Labrague LJ, Aguilar-Rosales R, Yboa BC, Sabio JB, de Los Santos JA. Student nurses' attitudes, perceived utilization, and intention to adopt artificial intelligence (AI) technology in nursing practice: a cross-sectional study. Nurse Educ Pract. 2023;73:103815. https://doi.org/10.1016/j.nepr.2023.103815
- 15. Sandanasamy S, McFarlane P, Okamoto Y, Couper AL. Knowledge and attitudes of nursing students towards artificial intelligence and related factors: a systematic review. J Nurs Rep Clin Pract. 2025;3(6):582-90. https://doi.org/10.32598/jnrcp.2408.1134
- 16. Swan BA. Assessing the knowledge and attitudes of registered nurses about artificial intelligence in nursing and health care. Nurs Econ. 2021;39(3):139-43. https://doi.org/10.62116/nec.2021.39.3.139
- 17. Shinners L, Grace S, Smith S, Stephens A, Aggar C. Exploring healthcare professionals’ perceptions of artificial intelligence: piloting the Shinners Artificial Intelligence Perception tool. Digit Health. 2022;8:20552076221078110. https://doi.org/10.1177/20552076221078110
- 18. Sommer D, Schmidbauer L, Wahl F. Nurses’ perceptions, experience and knowledge regarding artificial intelligence: results from a cross-sectional online survey in Germany. BMC Nurs. 2024;23(1):205. https://doi.org/10.1186/s12912-024-01884-2
- 19. Alruwaili MM, Abuadas FH, Alsadi M, Alruwaili AN, Elsayed Ramadan OM, Shaban M, et al. Exploring nurses' awareness and attitudes toward artificial intelligence: implications for nursing practice. Digit Health. 2024;10:20552076241271803. https://doi.org/10.1177/20552076241271803
- 20. Baraka AA, Alhijji MA, Alrajeh RA, Alotaibi DN, Bukhamis S, Alsafraa NA, et al. Acceptance and readiness of critical care nurses to use artificial intelligence: a structural equation modeling approach. Int Nurs Rev. 2025;72(2):e70035. https://doi.org/10.1111/inr.70035
- 21. Atalla AD, El-Gawad Mousa MA, Hashish EA, Elseesy NA, Abd El Kader Mohamed AI, Sobhi Mohamed SM. Embracing artificial intelligence in nursing: exploring the relationship between artificial intelligence-related attitudes, creative self-efficacy, and clinical reasoning competency among nurses. BMC Nurs. 2025;24(1):661. https://doi.org/10.1186/s12912-025-03306-3
- 22. Boyaci H, Soyuk S. Healthcare workers' readiness for artificial intelligence and organizational change: a quantitative study in a university hospital. BMC Health Serv Res. 2025;25(1):813. https://doi.org/10.1186/s12913-025-12846-y
- 23. Chen Y, Wu F, Zhang W, Xing W, Zhu Z, Huang Q, et al. Perspectives on AI-driven nursing science among nursing professionals from China: a qualitative study. Nurs Rep. 2025;15(6):218. https://doi.org/10.3390/nursrep15060218
- 24. Wei Q, Pan S, Liu X, Hong M, Nong C, Zhang W. The integration of AI in nursing: addressing current applications, challenges, and future directions. Front Med (Lausanne). 2025;12:1545420. https://doi.org/10.3389/fmed.2025.1545420
- 25. Karaca O, Caliskan SA, Demir K. Medical artificial intelligence readiness scale for medical students (MAIRS-MS) - development, validity and reliability study. BMC Med Educ. 2021;21(1):112. https://doi.org/10.1186/s12909-021-02546-6
- 26. Catalina QM, Fuster-Casanovas A, Vidal-Alaball J, Escale-Besa A, Marin-Gomez FX, Femenia J, et al. Knowledge and perception of primary care healthcare professionals on the use of artificial intelligence as a healthcare tool. Digit Health. 2023;9:20552076231180511. https://doi.org/10.1177/20552076231180511
- 27. Mousavi Baigi SF, Sarbaz M, Ghaddaripouri K, Ghaddaripouri M, Mousavi AS, Kimiafar K. Attitudes, knowledge, and skills towards artificial intelligence among healthcare students: a systematic review. Health Sci Rep. 2023;6(3):e1138. https://doi.org/10.1002/hsr2.1138
- 28. Davis FD. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 1989;13(3):319-40. https://doi.org/10.2307/249008
- 29. Venkatesh V, Morris MG, Davis GB, Davis FD. User acceptance of information technology: toward a unified view. MIS Q. 2003;27(3):425-78. https://doi.org/10.2307/30036540
- 30. Schepman A, Rodway P. Initial validation of the general attitudes towards artificial intelligence scale. Comput Hum Behav Rep. 2020;1:100014. https://doi.org/10.1016/j.chbr.2020.100014
- 31. Kim JM. Study on intention and attitude of using artificial intelligence technology in healthcare. J Converg Inf Technol. 2017;7(4):53-60. https://doi.org/10.22156/cs4smb.2017.7.4.053
- 32. Seo YH, Cho KA. Influence of AI knowledge, perception, and acceptance attitude on nursing students’ intention to use AI-based healthcare technologies. J Korean Nurs Res. 2022;6(3):81-90. https://doi.org/10.34089/jknr.2022.6.3.81
- 33. American Association of Colleges of Nursing. The essentials: core competencies for professional nursing education. Washington, DC: American Association of Colleges of Nursing; 2021.
- 34. Hong M, Shin H, Pi J. Artificial intelligence on nursing: a scoping review. J Converg Cult Technol. 2024;10(2):311-22. https://doi.org/10.17703/JCCT.2024.10.2.311
- 35. Lee H, Moon W, Kim S, Lee J, Zhang Y. Exploring the applicability of artificial intelligence for the improvement of nursing practice in Korea. J Korean Acad Nurs Adm. 2023;29(5):564-76. https://doi.org/10.11111/jkana.2023.29.5.564
E-SUBMISSION
