Factors Influencing Nursing Internship Students’ Readiness to Use AI: Cross-Sectional Study Using Neural Network Analysis

Abstract

Abstract Background Enhancing nursing students’ awareness, attitudes, beliefs, and preparedness toward AI may help improve their health care knowledge and practice. Objective This study aimed to assess nursing students’ attitudes, perceptions, self-efficacy, barriers, and anxiety, which influence their readiness to adopt AI in nursing practice. Methods This study used a cross-sectional, correlational design. Data were collected from 307 nursing internship students using an 8-part, self-administered questionnaire. Results Increased self-efficacy with computers was correlated with decreased barriers to accessing AI technology, lower computer anxiety scale scores (r=−0.27, P<.001 and r=−0.57, P<.001, respectively), and higher perceptions of using AI (r=0.27, P<.001). Meanwhile, nursing students’ readiness to adopt AI in nursing practice was negatively associated with barriers to accessing AI technology (r=−0.20, P<.001) and positively associated with attitudes toward and perceptions of using AI (r=0.32, P<.001 and r=0.14, P=.01, respectively). Increased barriers to accessing AI technology were associated with negative attitudes toward AI and nursing students’ perceptions of using AI (r=−0.34, P<.001 and r=−0.39, P<.001, respectively). A multilayer neural network model identified barriers (relative importance=0.27), attitudes (relative importance=0.16), and perceptions (relative importance=0.15) as the most significant predictors, while self-efficacy (relative importance=0.11) and anxiety (relative importance=0.07) showed smaller contributions, despite nonsignificant bivariate associations with nursing students’ AI readiness. The model demonstrated strong predictive performance, achieving a low relative error of 0.62 in the training set. The stability and generalization ability of the model were supported by the training and testing set results, which yielded a training sum of squares error of 65.93 and a testing sum of squares error of 35.49, showing no signs of overfitting. Conclusions Several contributing factors influenced nursing students’ readiness to embrace AI, with barriers, attitudes, and perceptions emerging as the most consistent, whereas self-efficacy and anxiety may play indirect roles. To improve the adoption of AI among nursing students, such factors should be dealt with in such educational programs; an interrelated adoption of AI in nursing practice is expounded as a more favorable environment.