Effectiveness of an artificial intelligence-assisted training program on cleaning competency among hospital environmental service staff

BACKGROUND: Healthcare-associated infections remain a major patient-safety threat, and environmental contamination is a key driver. Hospital cleaning staff are essential for prevention, yet conventional training lacks standardization and real-time feedback. We evaluated whether an artificial intelligence (AI) system with live monitoring and voice prompts improves terminal cleaning competency. METHODS: We randomly assigned 60 environmental service staff from Sir Run Run Shaw Hospital, Hangzhou, China, to either conventional instruction or AI-assisted training. The AI platform used computer vision and deep-learning algorithms to track cleaning tasks in real time against a predefined nine-item procedure and provided instant voice corrections when deviations occurred. Competency was assessed post-training under identical, feedback-free conditions, and a score of ≥ 7 was required to pass. RESULTS: Median competency scores were 7.8 (IQR: 7.4–8.4) in the AI group versus 5.8 (IQR: 5.4–7.4) in controls. All 30 AI-trained staff passed, compared with 9 (30%) controls. Largest gains occurred in edge cleaning, S-pattern mopping and high-touch surface disinfection. CONCLUSIONS: Real-time AI supervision significantly enhances cleaning competency and procedure adherence, offering a scalable and standardized approach to improve environmental hygiene and infection prevention.