Psychological factors associated with Chinese users’ adoption of AI-generated health education short videos: a PLS-SEM–ANN–IPMA study

This is the authors' abstract. We don't add key points for this paper.

Background AI-generated health education short videos are becoming an important form of digital health communication. Compared with conventional health education content, these videos require users to evaluate not only the value of health information, but also the credibility of information sources, the trustworthiness of AI-generated content, and their own health-related psychological state. Therefore, it is necessary to further examine the psychological factors that influence users’ adoption of AI-generated health information. Methods Drawing on the Information Adoption Model, this study developed a research framework that includes information quality, source credibility, perceived usefulness, AI trust, health anxiety, and information adoption intention. A questionnaire survey was conducted among Chinese users, yielding 476 valid responses. The data were analyzed using partial least squares structural equation modeling, artificial neural networks, and importance–performance map analysis. Results The PLS-SEM results showed that information quality and source credibility were significantly and positively associated with both perceived usefulness and AI trust. Perceived usefulness, AI trust, and health anxiety were also significantly and positively associated with information adoption intention. Among these factors, AI trust showed the strongest association with information adoption intention, whereas health anxiety was statistically significant but had a relatively weak effect. The mediation analysis further indicated that perceived usefulness and AI trust played important mediating roles between information-related factors and information adoption intention. The ANN results also identified AI trust as the most important predictor of information adoption intention. The IPMA results further showed that AI trust had the highest importance but a relatively insufficient performance level. Conclusion This study extends the Information Adoption Model to the context of AI-generated health education short videos and highlights the central role of AI trust in users’ information adoption process. The findings suggest that improving transparency, strengthening source credibility, providing evidence-based content, and introducing expert review mechanisms may help enhance users’ trust in AI-generated health information and further increase their information adoption intention.