Artificial Intelligence Adoption in Surgery: Cognition, Usage Patterns and Implementation Barriers of DeepSeek Among Healthcare Professionals in China’s Tertiary Hospitals

Hua Xie,1 Xu Dai,2 Jiao Xie,3 Shanshan Lei,2 Jie Zeng,2 Jiping Yang,1 Yifeng Zhou1 1Department of Operating, Hunan Provincial People’s Hospital (The First Affiliated Hospital of Hunan Normal University), Changsha, Hunan, 410005, People’s Republic of China; 2Department of Breast and Thyroid Surgery, Hunan Provincial People’s Hospital (The First Affiliated Hospital of Hunan Normal University), Changsha, Hunan, 410005, People’s Republic of China; 3Department of Otolaryngology, Second Xiangya Hospital of Central South University, Changsha, Hunan, 410011, People’s Republic of ChinaCorrespondence: Yifeng Zhou, Department of Surgery, Hunan Provincial People’s Hospital (The First Affiliated Hospital of Hunan Normal University), No. 61, Jiefang West Road, Furong District, Changsha, Hunan, 410005, People’s Republic of China, Tel +86 13974853659, Email [email protected] Jiping Yang, Department of Surgery, Hunan Provincial People’s Hospital (The First Affiliated Hospital of Hunan Normal University), No. 61, Jiefang West Road, Furong District, Changsha, Hunan, 410005, People’s Republic of China, Tel +86 15074962768, Email [email protected]: This study aims to investigate the cognition and application status of DeepSeek among surgical medical staff in Class III Grade A hospitals and analyse its influencing factors to optimise its clinical application.Methods: From February to March 2025, a questionnaire survey was conducted among 440 surgical medical staff from Class III Grade A hospitals in 18 provinces in China by a convenience sampling method. The questionnaire covered, among others, basic information, cognition and use status, attitudes and barriers and training needs, and also evaluated the degree of understanding, user experience (operation convenience, results accuracy, system stability) and degree of trust concerning DeepSeek. The SPSS 25.0 software was used for data analysis, which included frequency, percentage, Mann–Whitney U-test, Kruskal–Wallis H-test and multivariate ordinal logistic regression analysis.Results: A total of 424 valid questionnaires were collected (96.4%). The results indicated that 67.0% of the medical staff understood the basic functions of DeepSeek, and 70.3% used DeepSeek occasionally. It was mainly used for teaching and research support (43.2%), other life services (35.6%) and patient services (29.2%). Multivariate analysis showed that medical staff working in operating rooms and neurosurgery departments, those who were occasional users, and medical staff who primarily used DeepSeek for other life services demonstrated significantly higher levels of knowledge about DeepSeek.Conclusion: Despite widespread awareness of DeepSeek’s capabilities (67.0% understanding basic functions), significant implementation gaps persist, with limited clinical utilisation and predominant usage in low-risk applications. Key barriers include insufficient training (94.8% untrained), data privacy concerns (57.5%) and over-reliance fears (58.5%). These findings reveal a substantial untapped potential for AI integration in surgical practice, highlighting critical needs for targeted training interventions, enhanced data security frameworks and staged implementation protocols to bridge the awareness-utilisation gap and facilitate meaningful clinical adoption.Keywords: surgical medical staff, DeepSeek, knowledge of technology, behaviour of use, multi-centre survey