Content area
Aim
This study investigates the mediating role of artificial intelligence self-efficacy (AISE) in the association between AI literacy and learning engagemeg nursing students in higher education.
Background
The fourth technological revolution driven by AI has profound implications for healthcare education. Insufficient AI literacy and AISE among nursing students may hinder their learning engagement, yet empirical evidence on these relationships remains limited.
Design
A cross-sectional study.
Method
A convenience sampling method was used to select 2029 nursing students from 29 colleges in China from November to December 2024. A general information questionnaire, AI literacy scale, AISE scale and learning engagement scale were used to conduct the online questionnaire survey. Data were analyzed using descriptive analysis, Pearson correlation test and Bootstrap method.
Results
Results show that AI literacy was positively correlated with AISE and learning engagement ( r = 0.462–0.435, P＜0.01) and AISE was positively correlated with learning engagement ( r = 0.537, P＜0.01). The indirect effect of AI literacy on learning engagement through AISE was 0.130, accounting for 20.44 % of the total effect.
Conclusion
The AI literacy of nursing students is suboptimal; thus, enhancing foundational AI education is essential. AI literacy directly predicts learning engagement, with AISE serving as a partial mediator. Educators should integrate AI training into curricula, provide practical scenarios and foster interdisciplinary collaboration to improve AI competency and engagement.
1 Introduction
Artificial intelligence (AI) refers to the simulation of human intelligence in machines, performing tasks typically within the scope of human cognition, such as problem-solving, decision-making and pattern recognition ( Martinez-Ortigosa et al., 2023; O'Connor et al., 2023). The emergence and rapid development of AI have brought about disruptive changes in all fields of society, becoming an indispensable part of the healthcare field. It influences medical decision-making, nurse-patient relationships and the role of nurses, among other aspects, enhancing the accuracy of patient diagnoses and promoting the formulation of personalized treatment plans( Sharma and Jindal, 2024; Sheliemina, 2024).AI has great potential in the field of nursing education. For example, AI-innovated health educators can use data to provide students with personalized assessment and training ( Lomis et al., 2021). AI can virtually simulate scenes including medical emergencies, allowing nursing students to immerse in practice ( Glauberman et al., 2023). AI tutors can guide students to conduct patient interviews or provide immediate feedback on homework ( Sun and Hoelscher, 2023). With the rapid advancements in AI technology shaping the future of nursing education, nursing students must remain informed, adapt to innovations and use AI into their practice. This approach is vital to promote the recovery of patients and the development of the health care system ( Labrague et al., 2023).
1.1 Background
With the increasing digitalization of health care and the widespread popularity of electronic health data, AI and digital tools are transforming clinical nursing and enhancing interdisciplinary cooperation ( Abuzaid et al., 2022).AI literacy refers to an individual’s ability to understand, apply and evaluate AI applications and products to complete tasks ( Long and Magerko, 2020). It is a crucial competency in modern nursing education ( El-Sayed et al., 2024).Cultivating students’ AI literacy encourages them to apply AI technology in problem-solving, innovatively tackle nursing practice problems and cultivating their critical thinking ( Le Lagadec et al., 2024).
Nursing students are potential AI users and the backbone of future healthcare services. Their improvement in AI literacy not only plays a crucial role in leading and shaping the application and development of AI in the nursing field, but also affects their personal career development ( McGrow, 2019).Research shows that AI literacy boosts self-efficacy. Students excellent at using AI can perform several things. They can analyze complex data, work closely with teams from different fields, develop new solutions for clinical nursing problems and confidently showcase their skills and career potential. Thus, AI literacy is a key predictor of self-efficacy ( El-Sayed et al., 2024).
Self-efficacy is the belief in one’s ability to achieve expected goals and attain success, influencing people’s motivation and behavior persistence, as well as their attitude and effort in the face of difficulties ( Bandura, 1977; Jiang XiaoLian and GuoCheng, 2004).AI self-efficacy (AISE) refers to an individual’s confidence in successfully completing AI-related tasks ( Wang and Chuang, 2024). Research has shown that nurses with high AI self-efficacy are more open to medical AI and are more willing to apply new technologies to solve problems in clinical practice, thereby promoting innovative behavior ( Du et al., 2024; Rahman et al., 2016).For nursing students, robust self-efficacy can enhance their problem-solving abilities, strengthen their learning motivation and improve their academic performance ( Zhang et al., 2024).
Learning engagement is a proactive learning mindset, serving as the foundation for individuals to exert their subjective initiative, achieve excellent academic results and reach a high level of technical proficiency, thereby influencing the overall learning atmosphere and professional development ( Junjian et al., 2015). The learning engagement of nursing students is crucial for their professional growth and career development. It not only enhances their understanding of nursing knowledge, but also allows them to practice problem-solving skills, cultivate critical thinking and strengthen their ability for self-directed learning ( Huang et al., 2023). Students with high self-efficacy demonstrate better focus and are more willing to take on challenges and overcome difficulties, improving the effectiveness of their learning engagement ( Yufan et al., 2024).
Currently, studies on AI literacy of college nursing students at home and abroad are scant ( Yajun et al., 2025). Moreover, the relationship among AI literacy, AISE and learning engagement remains unclear. This study aims to understand the connections among AI literacy, AISE and learning engagement in nursing students in higher education, providing a theoretical basis for the reform of nursing education and offering references for educators to develop teaching strategies that adapt to the changing times. Consequently, high-quality nursing students, who can adapt to future healthcare environments, are cultivated.
2 Participants and methods
2.1 Design and setting
A cross-sectional study was designed, with nursing students in 29 higher universities and colleges in 11 Provinces in China. This study was reported using the STROBE cross-sectional research statement.
2.2 Participants
A convenience sampling was used to select nursing students from 29 universities in China from November to December 2024. Inclusion criteria: (1) full-time nursing students with an associate degree, bachelor’s degree, or higher; (2) informed consent and voluntary participation in this study. Exclusion criteria: (1) those who are on leave of absence during the investigation period. According to the Kendall sample size calculation method ( Ping et al., 2010), the formula is n = [Max(number of items)× (5 −10)]× [1 + (10 %−30 %)]. In this study, considering the maximum number of items is 22, the multiplier is 10 and the sample dropout rate is 10 %, the required sample size is 242. This study was reviewed and approved by the Ethics Committee of Hunan Province Integrated Traditional Chinese and Western Medicine Hospital (Ethics Review No. 258, 2024).
2.3 Method
2.3.1 Instruments
2.3.1.1 General information questionnaire
Based on a literature review combined with group discussions, developed by the researchers’ team, it primarily includes age, gender, educational background, family residence, parents’ education level, total average monthly income of the family and whether the school offers AI courses/training.
2.3.1.2 Artificial intelligence literacy scale
The AI literacy scale was developed by Wang et al. ( Wang et al., 2023). It includes four dimensions: awareness (three items), usage (three items), evaluation (three items) and ethics (three items). The scale uses a 7-point Likert system, ranging from “Strongly disagree” to “Strongly agree,” assigned values from 1 to 7 in order. A higher score indicates higher AI literacy level of the individual. Confirmatory factor analysis showed that CFI= 0.99, TLI= 0.99, GFI= 0.98, RMSEA= 0.01, SRMR= 0.03, the Cronbach’s α coefficient of the complete scale was 0.830 and the Cronbach’s α coefficient in this study was 0.829.
2.3.1.3 Artificial intelligence self-efficacy scale
AISE scale, developed by Wang and Chuang (2024), was used to assess users’ self-efficacy when using AI technologies and products. The scale includes four dimensions: assistance (seven entries), anthropomorphic interaction (five entries), comfort with AI (six entries) and technical skills (four entries), totaling 22 entries. It is scored on a seven-point Likert scale, ranging from “Strongly Disagree” to “Strongly Agree,” with values assigned from 1 to 7. A higher score indicates higher AISE level for the user. Confirmatory factor an-alysis yielded χ 2/df= 1.984, CFI= 0.941, TLI= 0.930, RMSEA= 0.079 and SRMR= 0.071, with Cronbach’s α coefficients of 0.958 and 0.828 in this study.
2.3.1.4 Learning engagement scale
The scale was developed by Schaufeli et al. (2002) and revised by Laitan et al. (2008) as a quantitative tool for assessing an individual’s psychological state and behavioral performance in learning activities. It mainly includes three dimensions: vigor (six entries), dedication (five entries) and concentration (six entries), totaling 17 entries. It is scored on a seven-point Likert scale, ranging from 1 (never) to 7 (always/every day). A higher score indicates higher the level of student engagement in learning. Confirmatory factor analysis revealed that χ 2/df= 2.77, CFI= 0.930, TLI= 0.920 and RMSEA= 0.080, with Cronbach’s α coefficients of 0.951 and 0.878 in this study.
2.3.2 Data collection
This study conducted an online questionnaire survey using Questionnaire Star. The researchers explained the purpose, significance and precautions of the study to the teaching heads of nursing colleges at the universities. After obtaining their informed consent, the university heads were asked to forward the questionnaire poster to class WeChat groups or the learning platform. An introductory statement was included at the beginning of the questionnaire to introduce the significance of the study to the nursing students participating in the survey, using uniform language. To ensure data quality, the researchers incorporated general knowledge questions and set a minimum response time of three minutes. Ultimately, 36 invalid questionnaires were screened out. A total of 2029 valid questionnaires were collected, with a recovery rate of 98.25 %.
2.3.3 Statistical analysis
SPSS 25.0 statistical software and Process plug-in (version 3.4) were used for data analysis. Frequency and percentage were used to express count variables in the general data questionnaire, whereas mean±standard deviation represented variable scores. The Pearson method was used to analyze the correlation among AI literacy, AISE and learning engagement of college nursing students. The mediating effect of AI literacy, AISE and learning engagement was verified using the Bootstrap method with the Process plug-in. A difference was considered statistically significant at P＜ 0.05.
3 Results
3.1 General information of college nursing students
In this study, 289 (14.2 %) were male and 1740 (85.8 %) were female; 1858 (91.6 %) were ＜22 years old, 161 (7.9 %) were 23–26 years old and 10 (0.5 %) were ＞27 years old; 1331(65.6 %) were junior college students, 528(26 %)were undergraduate students and 170 (8.4 %) were graduate students. Demographic details are summarized in
3.2 Scores of AI literacy, AISE and learning engagement among college nursing students
The results of this study showed that the total AI literacy score among nursing students in colleges was 44.97 (SD 9.83), indicating a below-average level. Scores in the ethics and awareness dimensions were similar, whereas the evaluation dimension scored the lowest. The total AISE score was 94.35 (SD 22.13), reflecting an above-average level, with the lowest score in technical skills and the highest in the help dimension. The total learning engagement score was 77.59 (SD 20.23), also above average, with the lowest score in vigor and the highest in dedication, as shown in
3.3 Correlation of AI literacy, AISE and learning engagement of college nursing students
The results of the Pearson correlation analysis demonstrate that AI literacy and its dimensions among university nursing students are positively correlated with AISE and its dimensions (
r = 0.325–0.468,
P < 0.05) and with learning engagement and its dimensions (
r = 0.402–0.435,
P < 0.05). In addition, AISE and its dimensions are positively correlated with learning engagement and its dimensions (
r = 0.501–0.537,
P < 0.05), as shown in
3.4 Mediation effect of AISE between AI literacy and learning engagement of college nursing students
A structural equation model is established, considering learning engagement as the dependent variable, AI literacy as the independent variable and AISE as the intermediary variable, as shown in
4 Discussion
This study found that nursing students in Chinese colleges scored an average of 44.97 ± 9.83 on AI literacy measures, reflecting a slightly below-average level. Scores in the ethics and awareness dimensions were similar, whereas the evaluation dimension was the lowest. Several factors may contribute to these findings. First, nursing curricula generally lack sufficient content related to intelligent technologies. A survey question “Does your school offer AI-related courses or lectures?” revealed that only 33.22 % of nursing students responded “Yes,” indicating that nursing education in China is still in its infancy. Second, the exam-oriented education system in the country tends to inhibit students’ evaluative and innovative abilities ( Loyalka et al., 2021). Third, many instructors lack experience in teaching AI-related topics, causing difficulty to provide effective instruction and support. Although ethics education receives some emphasis, students’ understanding of ethical and privacy issues often remains superficial and theoretical, with few opportunities for real-world application. To address these gaps, curriculum design must be enhanced, more practical training must be integrated, faculty capacity must be established and case-based and scenario-driven approaches in ethics education must be adopted ( Ng et al., 2021) ( Yao et al., 2024).
AI literacy was significantly and positively correlated with learning engagement among college nursing students, consistent with previous findings ( Fan and Zhang, 2024). Intelligent tutoring systems and adaptive learning platforms powered by AI offer personalization, dynamic learning experiences and emotional responsiveness, allowing students to adjust their learning pace based on their progress. When they encounter difficulties in learning, timely support and feedback can enhance their participation and promote learning engagement ( Gupta et al., 2024; Halkiopoulos and Gkintoni, 2024).In addition, in the digital age, AI technology bridges the gap between theoretical knowledge and practical application by simulating real scenarios, allowing them to refine their practical skills in a structured and responsive environment ( Buchanan et al., 2021; Montejo et al., 2024). For example, AI chatbots can simulate patient visits, allowing students to evaluate patients with health problems and develop nursing care plans ( Tam et al., 2023). This approach enhances their AI literacy, indirectly stimulates their interest in AI technology and promotes learning engagement ( Qiong et al., 2024).
Higher education nursing students’ AISE was significantly and positively associated with learning engagement, consistent with previous research findings ( Yufan et al., 2024).In line with self-efficacy theory, an individual’s self-efficacy influences their learning motivation and behavior ( WenXia and Guiping, 2006).Students with high self-efficacy exhibit strong confidence in their ability to successfully complete academic tasks. They are willing to take on challenging assignments, demonstrate resilience under high-intensity learning pressures and invest greater effort. This heightened engagement enhances their academic performance and ultimately supports the completion of their tasks ( ChunYing et al., 2023).
AI literacy among college nursing students is positively correlated with AISE. An individuals’ familiarity with technology influences their self-efficacy ( Falebita and Kok, 2025). Students with higher AI literacy usually have a deeper understanding of AI technology. When students continuously engage with and use AI-related tools in their learning process, their self-efficacy increases with the number of practices, increasing their willingness to take on challenging tasks and fostering a virtuous cycle.
The analysis in this study found that the indirect effect of AI literacy on learning input through AISE was 0.130, accounting for 20.44 % of the total effect. AISE partially and significantly mediated the relationship between AI literacy and learning input, suggesting that AI literacy could not only directly and positively influence the learning input behavior of college nursing students, but also partially mediate the role of AISE on learning input through college nursing students’ positive engagement. Individuals with high self-efficacy are confident in completing their learning tasks, developing clear completion plans, seeking solutions during implementation and continuously adjust their learning strategies, thereby promoting learning engagement ( Weiye and Jufang, 2021).Furthermore, based on the theory of knowledge, belief and action, behavioral change requires knowledge accumulation, followed by attitude formation and transformation, ultimately leading to behavioral shifts( Salas-Zapata et al., 2018). College nursing students acquire AI-related knowledge and application skills through various educational and training methods, including self-study, leading to their foundational understanding of AI and a certain level of AI literacy. Throughout the learning process, trust and understanding of AI technology are promoted, enhancing AISE. AISE improvement could stimulate nursing students’ enthusiasm for learning, increasing their willingness to invest more time and energy in practice. Consequently, they effectively use AI technology to improve their efficiency and quality of learning.
This study has certain limitations. First, as a cross-sectional study, it does not reflect the long-term psychological dynamics of AI literacy, AISE and learning engagement among nursing students in higher education. Second, the use of convenience sampling may introduce some selection bias, despite the implementation of uniform guiding language and a strict data screening process. Finally, the nursing students participating in this study come from different regions and not all schools have prioritized AI education or offered AI-related courses. Future research could address these limitations by conducting longitudinal studies using stratified sampling methods to explore the mechanisms and trajectories of dynamic changes in AI literacy, AISE and learning engagement. Moreover, a mixed-methods approach could be used in the study design.
5 Conclusion
Higher education nursing students’ AI literacy falls within the lower–middle range, whereas AISE and learning engagement are in the upper–middle range, with potential for further improvement. AISE serves as a partial mediator between AI literacy and learning investment among college nursing students.AI has a promising future in medical education and nursing educators must recognize the transition from a traditional teacher-student model to a teacher-machine-student interaction model ( Liming et al., 2024). AI-related courses can be categorized and graded to facilitate the transformation and efficient use of digital resources, providing an optimal teaching environment for educators and students. At the same time, interdisciplinary team teachers should be encouraged to collaborate through multi-team, in-depth cooperation to promote the development of AI teaching application tools ( Fei et al., 2024). Teachers must continuously update their AI-related knowledge and skills, expand their expertise in AI literacy and align teaching strategies with personalized learning requirements. The teaching program is designed based on the AI platform to strengthen the AI basic education of college nursing students, improve their AI literacy and AISE and promote the learning engagement of college nursing students.
CRediT authorship contribution statement
Sha Gong: Writing – review & editing, Writing – original draft, Investigation, Data curation. Shaoli Fu: Writing – review & editing, Writing – original draft, Investigation, Data curation. Fan Zhang: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Data curation. Jia Liu: Writing – review & editing, Writing – original draft, Investigation, Data curation. Lijuan Wang: Writing – review & editing, Writing – original draft, Supervision, Investigation, Data curation. Yongheng Li: Writing – review & editing, Supervision, Methodology, Conceptualization. Jing Huang: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Conceptualization. Taotao He: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation.
Declaration of Competing Interest
We declare that we have no conflicts of interest related to the research and publication of the manuscript titled " The mediation effect of AI self-efficacy between AI literacy and learning engagement in college nursing students: a cross-sectional study" submitted to Nurse Education in Practice.
We have received no financial support or funding that could have influenced our research or biased our findings. We have no personal or professional relationships that might pose a conflict of interest with regards to this work.Our research was conducted with the utmost integrity and adherence to ethical standards.
We hereby affirm that the contents of this manuscript represent our original work and have not been previously published elsewhere. We hold ourselves accountable for the accuracy and integrity of the data presented in the paper, and we are committed to the transparent reporting of our findings.
Acknowledgments
The authors thank all the nurse student who participated in this survey. Besides, the authors are grateful to all team staff.
Table 1
| Variables | Categories | N（%） |
| Age | ＜22years old | 1858（91.6 %） |
| 23–26years old | 161（7.90 %） | |
| ＞27years old | 10（0.50 %） | |
| Gender | Male | 289（14.2 %） |
| Female | 1740（85.8 %） | |
| Education | Postgraduate | 170（8.40 %） |
| Undergraduate | 528（26 %） | |
| Junior college | 1331（65.6 %） | |
| Family residence | Countryside | 1371（67.60 %） |
| Town | 658（32.40 %） | |
| Only child | Yes | 297（14.60 %） |
| No | 1732（85.4 %） | |
| Parents' education | Middle school and below | 1092（53.8 %） |
| Technical secondary school/high school | 737（36.3 %） | |
| Junior college | 120（5.90 %） | |
| Bachelor degree or above | 80（3.90 %） | |
| Total average monthly household income | ＜5000 yuan | 852（42 %） |
| 5000 ∼10,000 yuan | 911（44.90 %） | |
| ＞10,000 yuan | 266（13.10 %） | |
| Average monthly living expenditure | ＜1500 yuan | 1379（68 %） |
| 1501–2500 yuan | 575（28.30 %） | |
| ＞2501 yuan | 75（3.70 %） |
Table 2
| Dimension | Number of entries | Min-Max | Dimension score |
| AI literacy | 12 | 1～84 | 44.97 ± 9.83 |
| Awareness | 3 | 1～21 | 11.51 ± 2.60 |
| Usage | 3 | 1～21 | 11.44 ± 2.57 |
| Evaluation | 3 | 1～21 | 10.49 ± 3.59 |
| Ethics | 3 | 1～21 | 11.52 ± 2.71 |
| AISE | 22 | 1～154 | 94.35 ± 22.13 |
| Assistance | 7 | 1～49 | 32.43 ± 8.53 |
| Anthropomorphic interaction | 5 | 1～35 | 20.06 ± 6.00 |
| Comfort with AI | 6 | 1～42 | 25.89 ± 6.49 |
| Technical skills | 4 | 1～28 | 15.97 ± 4.58 |
| Learning Engagement | 17 | 1～119 | 77.59 ± 20.23 |
| Vigor | 6 | 1～42 | 26.28 ± 7.47 |
| Dedication | 5 | 1～35 | 23.68 ± 6.30 |
| Concentration | 6 | 1～42 | 27.63 ± 7.39 |
Table 3
| Variable | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 |
| 1Awareness | 1 | |||||||||||||
| 2Usage | 0.715** | 1 | ||||||||||||
| 3Evaluation | 0.592** | 0.700** | 1 | |||||||||||
| 4Ethics | 0.623** | 0.624** | 0.616** | 1 | ||||||||||
| 5AI literacy | 0.839** | 0.878** | 0.875** | 0.829** | 1 | |||||||||
| 6Assistance | 0.360** | 0.352** | 0.123** | 0.337** | 0.325** | 1 | ||||||||
| 7Anthropomorphic interaction | 0.399** | 0.401** | 0.283** | 0.389** | 0.421** | 0.580** | 1 | |||||||
| 8Comfort with AI | 0.431** | 0.433** | 0.234** | 0.425** | 0.430** | 0.728** | 0.766** | 1 | ||||||
| 9Technical skills | 0.440** | 0.437** | 0.327** | 0.426** | 0.468** | 0.473** | 0.702** | 0.723** | 1 | |||||
| 10AISE | 0.464** | 0.462** | 0.260** | 0.448** | 0.462** | 0.854** | 0.865** | 0.931** | 0.791** | 1 | ||||
| 11Vigor | 0.404** | 0.407** | 0.262** | 0.391** | 0.417** | 0.435** | 0.399** | 0.470** | 0.422** | 0.501** | 1 | |||
| 12Dedication | 0.398** | 0.391** | 0.234** | 0.396** | 0.402** | 0.496** | 0.381** | 0.491** | 0.386** | 0.518** | 0.859** | 1 | ||
| 13Concentration | 0.410** | 0.418** | 0.262** | 0.408** | 0.426** | 0.475** | 0.407** | 0.498** | 0.403** | 0.522** | 0.859** | 0.893** | 1 | |
| 14Learning Engagement | 0.423** | 0.425** | 0.266** | 0.417** | 0.435** | 0.489** | 0.415** | 0.509** | 0.423** | 0.537** | 0.951** | 0.955** | 0.961** | 1 |
| ** P＜0.05 |
Table 4
| Variables | Model path | Effect value | 95 % confidence interval | P | |
| LLCI | ULCI | ||||
| Direct effects | AI literacy→learning engagement | 0.345 | 0.287 | 0.403 | ＜0.001 |
| AI literacy→AISE | 0.376 | 0.345 | 0.408 | ＜0.001 | |
| AISE→learning engagement | 0.506 | 0.459 | 0.553 | ＜0.001 | |
| Indirect effects | AI literacy→AISE→learning engagement | 0.130 | 0.098 | 0.134 | ＜0.001 |
| Total effects | AI literacy→AISE→learning engagement | 0.636 | 0.592 | 0.679 | ＜0.001 |
© 2025 Elsevier Ltd