Research report: Artificial intelligence literacy and innovative behavior
A survey of 850 clinical nurses from several hospitals found that nurses who knew more about artificial intelligence (AI), felt more confident using it, and worked in places that supported new ideas tended to show more innovative behavior. The study took place at six large hospitals in Yueyang, Hunan, China. It also found that confidence in using AI statistically explained part of the link between AI knowledge and innovation. A workplace that encouraged new ideas made the link between confidence and innovative behavior stronger. However, these findings show relationships between factors, not proof that AI training, confidence, or workplace support caused nurses to become more innovative.
The researchers defined innovative behavior in nursing as a process in which nurses come up with new ideas, gain support for them, and put them into practice to solve clinical problems and meet patients’ needs. They used a 10-item questionnaire to measure three parts of this process: coming up with ideas, getting support, and putting ideas into action. The average score was 37.93 out of 50, with a standard deviation of 6.50.
The researchers carried out an online survey from June to July 2026. They used a sampling method that divided nurses into groups before selecting participants. Eligible participants were registered nurses working in clinical roles who had at least one year of clinical experience. They worked at two tertiary hospitals and four Grade-A tertiary hospitals in Yueyang. Four of the hospitals were general hospitals, and two specialized in particular types of care.
The researchers divided nursing units into internal medicine, surgery, and other groups. They aimed to recruit at least 40 nurses from each group at every hospital. The team collected 892 questionnaires and removed 42 that did not pass quality checks. Reasons for exclusion included finishing the questionnaire in less than five minutes, giving demographic answers that did not make sense together, or choosing the same answer for all 65 scale items. This left 850 valid responses, giving a valid-response rate of 95.3%.
A 12-item questionnaire with seven response levels measured AI literacy, including knowledge, use, evaluation, and ethical issues. Higher scores meant greater AI literacy. A separate 22-item questionnaire measured confidence in using AI, covering how AI can provide help, human-like interaction with AI, comfort with AI, and technical skills. The researchers measured the workplace climate for innovation using 21 items about rewards for new ideas, available resources, and management practices.
The 10-item Nurse Innovative Behavior Scale measured how nurses developed ideas, sought support, and put ideas into practice. The study reported Cronbach’s alpha values of 0.90 for AI literacy, 0.96 for confidence in using AI, 0.92 for workplace innovation climate, and 0.87 for innovative behavior. These values measure how consistently the items in each scale work together. A confirmatory factor analysis showed that the measurement model fit the data well. Its comparative fit index was 0.978, Tucker–Lewis index was 0.971, root mean square error of approximation was 0.063, and standardized root mean square residual was 0.028.
All four main factors were positively related to one another. Their correlation values ranged from 0.609 to 0.797, with P values below 0.01. AI literacy and confidence in using AI had the strongest relationship, with a correlation of 0.797. The relationship between workplace innovation climate and innovative behavior was 0.726. A multiple linear regression analysis found that education level, professional title, participation in research or innovation activities, AI literacy, confidence in using AI, and workplace innovation climate were associated with innovative behavior.
Together, these factors explained 60.9% of the differences in nurses’ innovative-behavior scores. The adjusted R-squared was 0.602. Workplace innovation climate had the largest standardized regression coefficient, at 0.502. Confidence in using AI followed at 0.211, and AI literacy was 0.124. These numbers describe how the factors related to innovative behavior in the statistical model. They do not prove that any one factor directly caused the behavior.
The researchers used a statistical mediation analysis to examine whether confidence in using AI helped explain the relationship between AI literacy and innovative behavior. Nurses with greater AI literacy tended to report greater confidence in using AI, and nurses with greater confidence tended to report more innovative behavior. When the analysis included confidence in using AI, the direct relationship between AI literacy and innovative behavior remained statistically significant. This suggests that confidence explained part, but not all, of the relationship.
The estimated indirect effect was 0.322, with a bootstrap 95% confidence interval from 0.243 to 0.403. The direct effect was 0.287, with a 95% confidence interval from 0.203 to 0.372. The indirect path made up 52.90% of the model’s estimated total effect. Because all measures were collected at the same time, these results describe a statistical relationship. They do not prove that learning more about AI increases confidence, which then leads to innovation.
The researchers also tested whether a workplace that supports innovation changed the strength of the relationship between confidence in using AI and innovative behavior. The interaction between confidence and workplace innovation climate was positive, with a coefficient of 0.051 and P < 0.01. An analysis of the relationship at different levels of workplace support found that confidence was positively related to innovative behavior in both less supportive and more supportive workplaces. However, the coefficient increased from 0.151 in workplaces with lower innovation-climate scores to 0.253 in workplaces with higher scores.
The estimated indirect effects through confidence in using AI were also larger when workplace innovation climate scores were higher. The effects were 0.120 at one standard deviation below the average climate score, 0.161 at the average, and 0.202 at one standard deviation above the average. None of the bootstrap 95% confidence intervals included zero. The index of moderated mediation was 0.041, with a bootstrap 95% confidence interval from 0.009 to 0.078. This means the estimated indirect relationship differed depending on how supportive the workplace was toward innovation.
The researchers based their analysis on the Ability–Motivation–Opportunity framework. This framework suggests that behavior is related to a person’s abilities, beliefs that affect motivation, and the opportunities available in their surroundings. In this study, AI literacy represented basic knowledge and ability. Confidence in using AI reflected how sure nurses felt about learning, using, and adapting to AI. Workplace innovation climate represented the support, resources, rewards, and practical opportunities they believed were available. The findings were consistent with the idea that nurses’ innovative behavior was related both to their own AI skills and confidence and to the conditions in which they worked.
The results do not show that personal confidence alone is enough to encourage innovation. The stronger relationship found in more supportive workplaces suggests that the link between confidence and behavior depended partly on workplace conditions. Comparisons made before adjusting for other factors also found differences in innovative-behavior scores by age, education level, years of work experience, professional title, marital status, type of work, AI-related training, frequency of AI-tool use, and participation in nursing research or innovation activities.
In the analysis that considered several factors at once, nurses with bachelor’s or master’s degrees, nurses with higher professional titles, and nurses who had taken part in or led innovation activities had higher innovative-behavior scores than the relevant comparison groups. Among the participants, 280 had attended AI-related training, 420 reported using AI tools one to three times per week, and 202 reported using them daily. Whether AI systems had been introduced at the hospital was not significantly related to innovative behavior in the initial analysis. However, this comparison did not establish whether such a relationship might exist in more complex statistical models.
Several limitations affect how the results should be understood. Because the study measured all factors at the same time, it cannot establish which came first or whether one factor caused another. The main measures were based on nurses’ own reports, so reporting bias is possible. However, an initial test for common-method bias did not find clear evidence of a major single-factor problem.
The mediation and moderated-mediation analyses did not adjust for demographic factors. The researchers also did not account for the possibility that nurses working at the same hospital might be more alike than nurses working at different hospitals. Although the sample included both general and specialized tertiary hospitals, all were in one Chinese city and all were tertiary institutions. This limits how well the findings may apply to other regions and smaller or lower-level hospitals. Future studies could combine nurses’ questionnaire responses with assessments from colleagues and head nurses. They could also examine objective measures, such as completed innovation projects, patents, quality-improvement results, and records showing how AI tools were used.
Long-term studies involving multiple centers and studies that test specific interventions are needed to find out whether improvements in AI skills, confidence, or workplace support lead to lasting increases in nurses’ innovative behavior.
Subject of Research: Relationships among AI literacy, confidence in using AI, workplace innovation climate, and innovative behavior among clinical nurses
Article Title: Artificial intelligence literacy and innovative behavior among clinical nurses: the roles of AI self-efficacy and organizational innovation climate
Article References: Li, J., Liu, X., Cao, Y., Tang, C., & Zhou, J. (2026). Artificial intelligence literacy and innovative behavior among clinical nurses: the roles of AI self-efficacy and organizational innovation climate. Frontiers in Public Health, 14, Article 1958854. https://doi.org/10.3389/fpubh.2026.1958854
Image Credits: AI Generated
DOI: 10.3389/fpubh.2026.1958854
Keywords: clinical nursing, artificial intelligence, AI literacy, confidence in using AI, nurse innovation, workplace climate, cross-sectional survey, tertiary hospitals, China, artificial intelligence literacy