Toward a human-centered digital ecosystem: bridging the "AI Chasm" in intelligent nosocomial infection monitoring.

Healthcare-associated infections (HAIs) pose a persistent global threat to patient safety, rendering traditional, resource-intensive manual surveillance increasingly unsustainable. While intelligent nosocomial infection monitoring systems (INIMS) offer a transformative digital solution, their clinical integration is frequently derailed by the "AI Chasm"-a critical misalignment between algorithmic design and clinical ergonomics. This Perspective transcends theoretical performance metrics to dissect real-world socio-technical barriers, including alert fatigue, dataset shift, and medico-legal liability. Unlike traditional socio-technical models (such as the SEIPS or NASSS frameworks) that primarily focus on the retrospective evaluation of health IT usability, our framework provides a proactive, pre-deployment governance strategy. To bridge this implementation gap, we propose a human-centered governance ecosystem anchored in two practical frameworks: "silent-mode validation" for safe, prospective clinical debugging, and "data solidarity" to ensure equitable resource allocation, particularly in low- and middle-income countries (LMICs). Ultimately, this framework outlines an actionable pathway to translate isolated technological innovations into sustainable, clinically integrated public health tools.