Profiles of artificial intelligence literacy and associations with evidence-based practice competence among nurses: a latent profile analysis Authors: Binmi Tang, Yali He, Yuxia Xiang, Yufang Chen Published: 2026 ABSTRACT AbstractAimsTo understand the current state of nurses’ artificial intelligence (AI) literacy. This study employs latent profile analysis to examine the relationship between different profile categories of AI literacy and evidence-based practice competence (EBPC) among nurses.MethodsFrom January to February 2026, nurses from Qingbaijiang District in Sichuan Province were selected by a self-designed general information questionnaire, the Artificial Intelligence Literacy Scale and the Questionnaire to Evaluate the Competency in Evidence-Based Practice of Registered Nurses. Latent profile analysis was performed to explore the profile categories of nurses’ AI literacy, the single-factor analysis and multivariate logistic regression analysis were employed to investigate the relevant influencing factors.ResultsThe AI literacy of nurses could be divided into three categories: the low AI literacy group (48.8%), the moderate AI literacy group (37.1%) and the high AI literacy group (14.1%). AI training and EBPC were the influencing factors of different profile categories (P < 0.001). These profile categories had significant effects on the nurses’ EBPC, as well as its four dimensions: attitude, knowledge, skill and application.ConclusionNursing administrators should effectively identify nurses with low AI literacy and develop personalized intervention plans to promote the development of AI literacy and EBPC.Clinical trial numberNot applicable.Keywords: Artificial intelligence literacy, Evidence-based practice, Latent profile analysis, Nurses, Nurse administrators Introduction The rapid development of artificial intelligence (AI) is reshaping traditional medical work models, with its application scope expanding from assisting disease diagnosis and providing personalized treatment plans to home robotic care [1]. In nursing practice, AI plays a significant role in optimizing patient care processes, improving patient health outcomes, enhancing nursing work efficiency, and reducing work burden [2–4]. In order to better cope with the wave of AI impact, the World Health Organization (WHO) has put forward new capability requirements for medical staff [5]. AI literacy is increasingly recognized as a crucial competency that nurses need to possess [6]. AI literacy refers to the comprehensive competence required for nursing professionals (including educators, students, and clinical practitioners) to correctly understand and evaluate AI technologies, make sound ethical judgments, and engage in innovative human-AI collaboration within an AI-driven healthcare environment [7, 8]. Kahraman et al. [9] reported that operating room nurses’ AI literacy was at a low-to-moderate level, with limited practical application of AI. Similarly, Ronquillo et al. [10] identified that nurses’ insufficient understanding of AI concepts and application contexts constrain effective AI use in nursing practice. Although nurses generally hold positive attitudes toward AI, their actual use of AI technologies remains limited [11]. Previous studies further indicate that nurses exhibit inadequate technical competence in applying AI tools, as well as limited awareness and capacity to address ethical concerns associated with AI applications [12]. The level of AI literacy has critical implications for nursing practice. Higher AI literacy enhances nurses’ ability to apply AI tools, thereby improving patient care quality [13, 14]. Conversely, low AI literacy may lead to misuse of AI tools, misinterpretation of AI-generated outputs, and erosion of patient trust [13]. However, current AI literacy assessment scales tend to focus mainly on technical operation, with limited exploration of ethical and cognitive dimensions [9]. This indicates that nurses’ AI capabilities in ethical and cognitive dimensions have not received sufficient attention. Therefore, it is necessary to strengthen their AI literacy from more comprehensive dimensions to fully realize the potential of AI in nursing practice [15]. Evidence-based practice competence (EBPC) refers to the comprehensive ability of nurses to acquire, screen, and evaluate evidence-based knowledge, and apply it in specific contexts based on clinical environments and patient preferences to solve nursing problems [16]. EBPC is one of the core competencies of nurses, and an important foundation for achieving optimal clinical outcomes for patients at a reasonable cost [17]. With the advancement of technology, AI is increasingly being integrated into evidence-based practice (EBP) [18]. Natural language processing technology can significantly improve the efficiency of evidence retrieval and personalized abstract generation [19, 20], while generative AI can assist users in evaluating evidence quality and predicting patient compliance [21]. For instance, Kelly et al. [22] proposed a “data-algorithm-training-output” method grounded in EBP steps, demonstrating AI’s role in streamlining processes and evidence evaluation in radiology research. However, the application of AI in EBP still faces many challenges. AI may produce incorrect outputs due to its inability to recognize subtle differences in research design or incomplete information extraction [23]. In addition, the training databases of some AI tools are not entirely from medical databases, which may lead to phenomena such as fictitious information [24]. It is essential to examine how nurses’ AI literacy affects their ability to apply evidence-based knowledge, in order to safeguard patient safety, enhance care quality, and minimize errors in the use of AI-generated recommendations. Nevertheless, few studies have examined the heterogeneous patterns of AI literacy among nurses or their association with EBPC. Therefore, this study aims to identify latent profiles of AI literacy among nurses using latent profile analysis (LPA) and to explore the differences in EBPC across these profiles. Self-determination theory offers a useful framework for understanding how AI literacy drives EBPC. Self-determination theory posits that an individual’s intrinsic motivation and degree of behavioral engagement are influenced by the level of satisfaction of three fundamental psychological needs: autonomy, competence, and relatedness [25]. This belief plays a significant role in enhancing intrinsic motivation, influencing behavioral choices, and bolstering confidence [26]. In nursing practice, AI literacy enables nurses to effectively utilize AI tools for searching, filtering, and preliminary evaluation of evidence resources. And it also helps nurses build confidence in their ability to handle information when facing complex clinical problems. When AI helps nurses feel competent and capable of making autonomous decisions, they will be more proactive in carrying out and adhering to EBP. Current research often uses single correlation or regression analysis methods to analyze nurses’ AI literacy and EBPC. But these studies cannot identify differences between groups and are difficult to explore causal relationships. This study aimed to identify latent profiles of AI literacy and then examine the relationship between these profile categories and nurses’ EBPC. Methods Study design A descriptive, correlational and cross-sectional research. Participants and sample size A convenience sampling method was used to recruit 537 nurses from a district in Chengdu, Sichuan Province from January to February 2026, and all nurses agreed to participate in the study. The inclusion criteria were as follows: (1) holders of a valid nurse practicing certificate; (2) ≥ 1 year of nursing work experience; (3) voluntary written informed consent. The exclusion criteria were as follows: (1) nursing students, trainee or advanced study nurses; (2) nurses who were absent from duty for > 1 month. This study included 53 items in total, drawn from a self-designed general information questionnaire, the Artificial Intelligence Literacy Scale, and the Questionnaire to Evaluate the Competency in Evidence-Based Practice of Registered Nurses. Following the common practice for sample size estimation in cross-sectional studies, the sample size was set at 5 to 10 times the number of independent variables. With an additional allowance of 20% for invalid responses, the minimum required sample size was calculated to be 318 participants. Meanwhile, Nylund-Gibson et al. [27] recommended at least 300 participants for latent profile analysis to ensure model stability. The final sample comprised 537 nurses, meeting both requirements. Instruments A self-designed general information questionnaire The information collected in this questionnaire includes age, gender, highest education level, year of work experience, position, and AI training. Artificial Intelligence Literacy Scale (AILS) AILS was developed by Wang et al. [28] to evaluate the AI literacy of the general population, including four dimensions: awareness, use, assessment, and ethics, with a total of 12 items. The scale adopts the Likert 7-level scoring method, assigning 1–7 points from “strongly disagree” to “strongly agree”, with a total score of 12–84 points. The higher the score, the better the AI literacy. The previous validation studies in China have reported Cronbach’s α coefficients between 0.830 and 0.911 [28, 29]. In this study, the Cronbach’s α was 0.882, indicating good internal consistency. Questionnaire to evaluate the competency in evidence-based practice of registered nurses (EBP-COQ-Prof) The scale was developed by Ruzafa-Martínez et al. [16] and translated by Yue et al. [30]. The scale includes four dimensions: attitude, knowledge, skill, and application, with a total of 35 items. The scale adopts the Likert 5-point scoring method, with 1–5 points representing “completely disagree” to “completely agree”, and a total score of 35–175 points. The higher the score, the higher the nurses’ EBPC. According to earlier validation studies in China, the Cronbach’s α was 0.980 [30]. In the current study, the Cronbach’s α was 0.983, demonstrating reliable internal consistency. Ethical consideration This study strictly followed the Declaration of Helsinki. The study has been approved by the Ethics Committee of Traditional Chinese Medicine Hospital of Qingbaijiang District Chengdu (No. 2026-KY-001). All participants completed and submitted an electronic informed consent form prior to taking part in the study. Participation was voluntary, and all responses were anonymized to ensure confidentiality. Data collection The data collection period was from January to February 2026. Distribute electronic questionnaires through online platforms to collect information on nurses’ social demographics, AI literacy, and EBPC. All nurses voluntarily participated in this study and remained anonymous throughout the entire process. Researchers will exclude questionnaires that take less than 5 min to fill out or are illogical. Statistical analysis Using Mplus 8.3 software for latent profile analysis (LPA) to identify profile categories of nurses’ AI literacy. As dimension or total scores of the AILS may mask item-level heterogeneity in AI literacy, we performed item-level LPA to retain maximal measurement information and identify latent classes of nurses’ AI literacy. Model estimation started with a one-profile solution, after which the number of profiles was increased incrementally. The following indicators were used to determine the optimal number of profiles: Akaike information criterion (AIC), Bayesian information criterion (BIC), adjusted Bayesian information criteria (aBIC), Lo-Mendell-Rubin test (LMRT), Bootstrapped likelihood ratio test (BLRT), with lower values indicating better model fit. Entropy values closer to 1 indicate higher classification accuracy. The LMRT and BLRT were used to compare adjacent models. A statistically significant result (P < 0.05) meant that the model with k profiles was superior in fit to the model with k-1 profiles. To ensure interpretability and stability, we required each latent class to account for at least 5% of the total sample. In this study, profiles were named based on their relative mean scores on the AILS. Continuous variables were expressed as mean (M) ± standard deviation (SD), while categorical variables were expressed as frequency (n) and percentage (%). One-way ANOVA, chi-square test, and Kruskal-Wallis H test were used to analyze the relationship between sociodemographic data and AI literacy profile categories. Multivariate logistic regression analysis was performed to identify factors associated with latent profile membership of nurses’ AI literacy. The three profile as independent variables, the total score of nurses’ EBPC and its four dimensions as dependent variables. Multiple linear regression was used to analyze the relationship between AI literacy profile categories and EBPC. All statistical tests are two-sided tests, with P < 0.05 considered statistically significant. Results Latent profile analysis of nurses’ AI literacy Latent profile analysis was conducted based on 12 items of AILS. The fit indices for each latent class model are shown in Table 1. As the number of profiles increased, the AIC, BIC, and aBIC values consistently decreased. For the four-profile model, the AIC, BIC, and aBIC values showed smaller reductions, and the P value of LMRT was not significant. When the five-profile model was reached, the decreases in AIC, BIC, and aBIC values became even smaller. When the three-profile model was reached, the entropy value approached 1 (0.964), indicating high classification accuracy. The P values for both the LMRT and BLRT were significant (P < 0.05). Moreover, the smallest profile size exceeded 5% of the total sample, indicating good model stability. Thus, the three-profile model was chosen as the best-fit model. Table 1Model fit indices for latent profile analysisIndicationsProfile12345AIC18323.26115970.90914791.95614401.1214011.177BIC18426.12516129.49115006.25614671.13813446.913aBIC18349.94116012.04114847.53914471.15514095.663LMRT-<0.0010.0010.1240.009BLRT-<0.001<0.001<0.001<0.001Entropy-0.9720.9640.9690.976Profile size(%)100.0081.4/18.648.8/37.1/14.148.4/36.3/5.1/10.20.7/48.0/36.2/10.1/5.0 Model fit indices for latent profile analysis The latent profile structure of nurses’ AI literacy is shown in Fig. 1. Profile 1 comprised 262 nurses (48.8%) who scored low on all items, and was named the low AI literacy group. Profile 2 included 199 nurses (37.1%) who scored moderate on all items, and labeled the moderate AI literacy group. Profile 3 encompassed 76 nurses (14.1%) and termed the high AI literacy group. AI literacy of the outcomes in each profile. Note AILS: Artificial Intelligence Literacy Scale Univariate analysis of latent profiles of AI literacy A total of 562 nurses participated in the survey. Excluding incomplete questionnaires (3) and questionnaires with a response time of less than 5 min or illogical responses (22), 537 questionnaires were ultimately included with an effective response rate of 95.55%. The average age of participants was 31.50 (SD 6.83). The majority of participants were female 520 (96.8%), with a bachelor’s degree of 351 (65.4%) and an average year of work experience of 10.09 (SD 8.50). AI training and EBPC were influencing factors in the profile categories of nurses’ AI literacy (P < 0.05). Details are presented in Table 2. Table 2Demographic characteristics and univariate analysis (n = 537)VariablesTotalProfile 1(n = 262)Profile 2(n = 199)Profile 3(n = 76)Test statistics P N(%)Mean ± SDMean ± SDMean ± SDAge(year)31.50 ± 6.8330.91 ± 6.2332.36 ± 7.2821.28 ± 7.450.711b0.892Gender0.753a0.686 male17(3.2)7(2.7)8(4)2(2.6) female520(96.8)255(97.3)191(96)74(97.4)Highest education level3.367a0.762 Vocational school education5(0.9)2(0.8)2(1)1(1.3) Associate degree177(33)91(34.7)58(29.1)28(36.8) Bachelor’s degree351(65.4)168(64.1)137(68.8)46(60.5) Master’s degree4(0.7)1(0.4)1(1)1(1.3)Position1.018a0.985 None357(66.5)174(66.4)130(65.3)57(69.7) Nursing Team Leader47(8.8)21(8)19(9.5)7(9.2) Nurse Manager60(11.2)30(11.5)23(11.6)7(9.2) Others73(13.6)37(14.1)27(13.6)9(11.8)Year of work experience10.09 ± 8.509.731 ± 9.1510.68 ± 7.699.8 ± 8.250.770b0.827AI Training20.586a<0.001 Yes183(34.1)66(25.2)79(39.7)38(50) No354(65.9)196(74.8)120(60.3)38(50)EBPC score127.37 ± 27.75112.25 ± 20.48137.22 ± 26.92153.7 ± 20.125.090b<0.001Note: Profile 1: low AI literacy group; Profile 2: moderate AI literacy group; Profile 3: high AI literacy group; a χ2 value, b H value Demographic characteristics and univariate analysis (n = 537) Note: Profile 1: low AI literacy group; Profile 2: moderate AI literacy group; Profile 3: high AI literacy group; a χ2 value, b H value Multivariate logistic regression analysis of factors influencing AI literacy among nurses The latent profile categories of AI literacy were used as the dependent variable, variables with P < 0.05 in the univariate analysis were then included as independent variables in a multivariate logistic regression model. The results showed that EBPC and AI training were significant factors related to nurses’ AI literacy (P < 0.05). Detailed results are presented in Table 3. Table 3Multivariate logistic regression analysis of factors influencing AI literacy among nursesVariableProfile 2 vs. Profile 1Profile 3 vs. Profile 1BOR95%CI P BOR95%CI P AI Training: Yes0.7532.123[0.387–1.120]<0.0011.2173.377[0.480–1.953]0.001EBPC score0.0481.049[0.040–0.055]<0.0010.0731.076[0.057–0.088]< 0.001Note: Profile 1: low AI literacy group; Profile 2: moderate AI literacy group; Profile 3: high AI literacy group Multivariate logistic regression analysis of factors influencing AI literacy among nurses Note: Profile 1: low AI literacy group; Profile 2: moderate AI literacy group; Profile 3: high AI literacy group Association of the AI literacy profiles with EBPC The linear regression models between AI literacy profiles and EBPC scores are presented in Table 4. The results showed that there was a correlation between AI literacy profiles and the total score as well as each dimension of EBPC. Table 5Pairwise comparisons between each profileVariablesMean DifferenceProfile 1 vs.Profile 2Profile 1 vs. Profile 3Profile 2 vs. Profile 3Overall testAttitude−7.44***−11.65***−4.21***98.532***Knowledge−6.00***−11.68***−5.68***67.949***Skill−4.86***−6.91***−2.05**115.323***Application−6.66***−11.20***−4.54***112.091***EBPC score−24.96***−41.45***−16.48***125.324***Note: **P < 0.001, ***P < 0.001; Profile 1: low AI literacy group; Profile 2: moderate AI literacy group; Profile 3: high AI literacy group Pairwise comparisons between each profile Note: **P < 0.001, ***P < 0.001; Profile 1: low AI literacy group; Profile 2: moderate AI literacy group; Profile 3: high AI literacy group Table 4Multivariable linear regression analyses for the association of the AI literacy profiles with EBPCVariablesBβ95%CI P Attitude Profile 311.6490.468[9.745,13.554]<0.001Profile 27.4410.414[6.067,8.816]<0.001Profile 1RefAdjusted R-squared: 0.267 Knowledge Profile 311.6810.437[9.546,13.817]<0.001Profile 25.9990.311[4.458,7.541]<0.001Profile 1RefAdjusted R-squared: 0.200 Skill <0.001Profile 36.9130.474[5.824,8.005]< 0.001Profile 24.8610.462[4.074,5.648]Profile 1RefAdjusted R-squared: 0.299 Application Profile 311.2000.504[9.530,12.870]< 0.001Profile 26.6620.415[5.457,7.867]< 0.001Profile 1RefAdjusted R-squared: 0.293 Multivariable linear regression analyses for the association of the AI literacy profiles with EBPC Figure 2 shows the total scores of EBPC and its four dimensions (attitude, knowledge, skill, application) in three AI literacy profile categories. Profile 3 (high AI literacy) showed the highest score of total score (153.70 SD 20.12) and each dimension of EBPC (attitude: 45.71 SD 5.69; knowledge: 42.25 SD 7.52; skill: 26.46 SD 4.08; application: 39.28 SD 6.42). Profile 2 (moderate AI literacy: total scores of EBPC: 137.22 SD 26.69; attitude: 41.50 SD 7.74; knowledge: 36.57 SD 9.93; skill: 24.41 SD 4.75; application: 34.74 SD 7.64). Profile 1 (low AI literacy) showed the lowest score on the total score (112.25 SD 20.48) and each dimension of EBPC (attitude: 34.06 SD 7.65; knowledge: 30.57 SD 7.17; skill: 19.55 SD 3.90; application: 28.08 SD 5.56). The pairwise comparison results among the profile categories are shown in Table 5. The results indicate that significant differences (P < 0.01) exist among the three profile categories in terms of the total score of EBPC and its four dimensions. EBPC scores across latent profiles. Note. EBPC: evidence-based practice competence Discussion This study employed LPA to identify different profile categories of nurses’ AI literacy, then analyzed the relevant influencing factors of these profile categories, and explored the relationship between the AI literacy profile categories and nurses’ EBPC. The research results provide a theoretical basis for formulating targeted intervention strategies for nurses with different AI literacy profile categories, thereby effectively enhancing their AI literacy. The study identified three groups of nurses’ AI literacy through LPA: low AI literacy, moderate AI literacy, and high AI literacy. The group with low AI literacy accounts for the highest proportion, and it is particularly weak in the dimensions of assessment and use, which is consistent with the research conclusions of Zhu et al. [29]. The reason may be that there is still a certain gap between the current application rate and maturity of AI in nursing practice in China and that in developed countries, and nurses generally lack opportunities to actually apply AI tools. The lack of assessment and application experience may make it difficult for nurses to make accurate judgments when facing AI output results. For this group, it is recommended to implement gradual practical training: starting with low-risk scenarios (such as information retrieval) to help them build confidence in basic operations, and then gradually transitioning to high-risk clinical decision support scenarios (such as disease early warning). The overall score of the moderate AI literacy group in the AILS items is at a medium level, with a relatively low score in the ethical dimension. The reason may be that a relatively high proportion of nurses in this group have not participated in AI training, lacking systematic theoretical guidance, making it difficult for them to form a systematic understanding of AI. In addition, although nurses are aware of the data privacy and security issues involved in AI, they lack the ability to understand and analyze specific ethical issues such as informed consent and the division of responsibility [13]. Unlike the low AI literacy group, the moderate literacy group possesses a certain foundation in AI application, but their ethical competence lags behind. For this group, the focus of AI training should shift from “application” to “ethics”, thereby helping them build the ability to identify and solve ethical issues. The group with high AI literacy scores high on all items of AILS, but also faces the issue of low scores in the ethics dimension. This is primarily attributed to the current emphasis on enhancing application capabilities in AI education and training, while neglecting in-depth attention to issues such as ethics, privacy, and risks [31]. For this group, in-depth ethics training should be provided, covering topics such as informed consent, personal data protection, and algorithm transparency and interpretability [32, 33]. Nurses who have not received AI training have a lower level of AI literacy. This is mainly because AI technology has a high cognitive threshold, and training is an important way to acquire basic knowledge and application skills. Nurses who have not received AI training often lack basic knowledge and understanding of AI and its applications. Moreover, training can alleviate the technical anxiety of users, thereby fostering a positive attitude towards AI applications. Therefore, in order to further enhance nurses’ AI literacy, it is necessary to incorporate AI technology into nursing education and clinical training [34]. Firstly, universities should offer AI training courses to nursing students [35]. Since the improvement of AI literacy is a gradual process, it is necessary to carry out phased training for nursing students [6]. In the first year, the curriculum should focus on basic AI knowledge and applications to help students develop a foundational understanding of AI. In the second year, nursing students should receive instruction on the application of AI in nursing practice through integration with specific healthcare scenarios. In the third year, the curriculum should focus on the implementation and application of AI in clinical practice, preparing nursing students for the future integration of AI technologies into nursing care. Medical institutions should implement stratified AI training programs for nurses. Tailored training content should be provided for nurses with different levels of AI literacy. For nurses with low AI literacy, training should focus on AI concepts and basic operations. For those with moderate AI literacy, emphasis should be placed on the application of AI in nursing practice and outcome evaluation. For nurses with high AI literacy, training should focus on AI ethics and the evaluation of AI tools. EBPC is an important factor influencing nurses’ latent profile categories of AI literacy. Nurses with higher EBPC generally demonstrate stronger abilities in evidence appraisal, data interpretation, and research-informed clinical decision-making, and are therefore more likely to understand and accept the outputs of AI. The results of the multi-group comparisons showed that sufficient AI literacy was associated with a high level of EBPC. Previous studies on nursing students have shown that a high level of information literacy can help enhance their EBP levels [36]. In addition, research on clinical nurses has also found that personal information literacy is one of the important factors affecting the implementation of EBPC [37]. It should be noted that this study did not include several potential confounding factors, such as prior computer experience and departmental support for new technologies. These factors may simultaneously influence both AI literacy and EBPC. Therefore, future studies should incorporate these variables into the model for further validation. This study indicates that a high level of AI literacy can enhance nurses’ EBPC, thereby promoting the development of EBP [38]. Nursing managers should attach importance to the role of AI literacy in promoting the development of EBPC. The reliable EBP must be grounded in rigorous clinical evidence and specific application scenarios, and researchers must carefully scrutinize the results outputted by AI [18]. Moreover, this study suggests incorporating AI into nursing curricula and clinical training, while incorporating EBP into the training content. These measures will help improve nurses’ EBPC simultaneously [6]. In the context of rapid advances in artificial intelligence, clinical nurses are increasingly reliant on AI tools. However, excessive reliance may lead to several potential problems, including reduced critical thinking ability, decreased autonomous decision-making capacity, and increased overtrust in algorithmic outputs, thereby contributing to automation bias [39]. Therefore, it is necessary to develop a human-machine collaborative evidence-based practice model in which AI serves as a supportive tool rather than a replacement for nurses in clinical decision-making [18]. Limitations This study has the following limitations. Firstly, as this was a cross-sectional study, causal relationships and underlying mechanisms between variables could not be established. In addition, the dynamic nature of EBPC across attitude, knowledge, skill, and application dimensions could not be fully captured. Secondly, the use of self-reported questionnaires may introduce social desirability and self-assessment bias, potentially affecting data accuracy. Thirdly, although several covariates were included, potential confounders such as nurses’ prior computer experience, departmental support for new technologies, and professional title may not have been fully controlled, which may influence the interpretation of the observed associations. Future studies should adopt longitudinal or prospective designs to clarify causal relationships and dynamic changes between AI literacy and EBPC. The use of mixed-method approaches and objective assessments is recommended to reduce measurement bias. In addition, more comprehensive confounding factors should be considered, and multi-center studies with larger samples are needed to improve generalizability across different clinical settings. Conclusion In the context of AI reshaping nursing practice, the study explored the relationship between nurses’ AI literacy and EBPC based on latent profile analysis. The research results reveal the importance of AI literacy to the development of nurses’ EBPC, and identify profiles of AI literacy with different characteristics. The research findings further underscore the significance of incorporating AI into nursing education and clinical training. At the same time, we emphasize that nursing educators and nursing managers should proactively identify nurses with lower AI literacy and develop targeted educational courses and training programs. This study offers a new perspective for nursing education and clinical training, and future research is encouraged to further explore how these approaches can enhance nurses’ AI literacy. Acknowledgements All authors sincerely appreciate the nurses who participated in this study. We thank them for their time, active cooperation, and the profound insights they shared. Author contributions Binmi Tang was responsible for data analysis, interpretation of results, manuscript writing and editing, Yali He, Yuxia Xiang, and Yufang Chen performed literature review, data collection and manuscript proofreading. Funding No external funding. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Human ethics and consent to participate. This study strictly followed the Declaration of Helsinki. The study was approved by the Ethics Committee of Traditional Chinese Medicine Hospital of Qingbaijiang District Chengdu (No. 2026-KY-001). Informed consent was obtained electronically from all participants. Participants received an email with a link to the e-questionnaire. The first page of the questionnaire included a link to the explanatory document of the study. A consent checkbox was provided on the survey page, and only those who checked the box were able to proceed to the e-questionnaire. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.