JMIR Publications

ABSTRACT

Artificial intelligence in health care has largely been framed around prediction, automation, and generative assistance, yet nursing work remains underrepresented in how clinical intelligence is conceptualized and built. A major reason is that routine documentation and electronic health record (EHR) interaction behavior are still treated primarily as administrative burden rather than as meaningful traces of clinical work. This Viewpoint argues that medical futures studies should recognize nursing documentation as a documentation-native signal environment: one that reflects not only recordkeeping, but also surveillance intensity, workflow strain, care coordination, and emerging clinical risk. The argument is grounded in a growing body of work on documentation burden, audit-log modeling, workflow-aware clinical intelligence, and EHR-derived temporal signals, including work developed through the AIM-AHEAD CLINAQ Fellowship, IRB-exempt study #2026-027, SIIM-CAIMI25, and related Intensive Documentation Index manuscripts. The central claim is not that documentation should be intensified, but that future nursing AI should be burden-aware, workflow-aware, nurse-centered, equitable, and implementation-realistic. Rather than merely accelerating chart completion, next-generation systems should identify burden, detect strain, surface changes in surveillance behavior, and support safer decisions without increasing cognitive load. We propose documentation-native, workflow-aware nursing AI as a field-shaping agenda for medical futures studies: one that treats routine EHR behavior not as administrative exhaust, but as an underused source of operational and clinical intelligence. Such a shift would better align AI development with real nursing work, implementation constraints, and the goal of safer, more equitable care.