Donghui Ren, Yufang Qu, Hui Sun
Using Healthcare-Associated Infection (HCAI) surveillance as a representative model for high-burden administrative tasks, this Perspective argues that healthcare worker fatigue, sleep disruption, and poorly integrated digital systems threaten both occupational health and health-system resilience. Crucially, these problems should be addressed as failures of work-system design rather than deficits in individual resilience. We propose a Human-Centered Digital Ecosystem that places adequate staffing, circadian-aligned scheduling, protected recovery, and supportive organizational conditions before technological intervention. Within this ecosystem, artificial intelligence could potentially function as a task-specific "cognitive shield" by ranking, consolidating, or filtering low-value information before it reaches clinicians. However, AI's role should be strictly adjunctive; it is not a substitute for structural reforms such as increased hiring, fair compensation, and cultural shifts. AI-supported tools might be considered adjunctive and could be applied only for clearly defined and potentially avoidable administrative burdens after major organizational risks have been assessed and are being addressed in parallel. Their introduction should not delay, replace, or weaken investments in staffing, workload management, fair compensation, supportive leadership, and protected recovery. Implementation should proceed through staged validation, workforce co-design, continuous assessment of workload redistribution, and governance mechanisms protecting autonomy, equity, data security, and emotional privacy. We contend that digital interventions should be judged not only by technical accuracy or time saved, but also by their effects on after-hours work, cognitive workload, recovery, professional judgment, and the distribution of work across occupational groups. The purpose of digital transformation should be to remove avoidable work, not to compensate for preventable organizational deficiencies.