ARTIFICIAL INTELLIGENCE AND NURSING: CLINICAL APPLICATIONS, PATIENT SAFETY, HUMANIZATION, AND ETHICAL GOVERNANCE
TL;DR
The evidence indicates that AI can improve risk stratification, early recognition of clinical deterioration, medication-safety processes, documentation quality, and access to simulation-based education, and support a human-in-the-loop model in which AI augments rather than replaces nursing judgment.
Abstract
Artificial intelligence (AI) is increasingly reshaping nursing practice by supporting clinical decision-making, patient monitoring, documentation, education, workflow management, and patient safety. This integrative literature review critically examines the principal applications, benefits, limitations, and ethical implications of AI in nursing, with emphasis on emergency triage, hospital data management, clinical decision support, humanized care, and safety governance. A structured search of PubMed/MEDLINE and major peer-reviewed journal platforms was conducted, prioritizing literature published from 2020 onward while retaining seminal works required for the theoretical framework. Peer-reviewed original studies, systematic and integrative reviews, umbrella reviews, and authoritative institutional guidance were considered. The evidence indicates that AI can improve risk stratification, early recognition of clinical deterioration, medication-safety processes, documentation quality, and access to simulation-based education. Nevertheless, the literature remains heterogeneous, and many applications lack prospective validation across diverse health systems and populations. Major concerns include algorithmic bias, privacy, opacity, automation bias, alert fatigue, unequal technological access, and unclear accountability. The findings support a human-in-the-loop model in which AI augments rather than replaces nursing judgment. Sustainable adoption therefore requires local validation, continuous performance monitoring, AI literacy, interdisciplinary governance, and explicit protection of patient autonomy and relational care. For Brazil, further multicenter research is needed to evaluate AI under the organizational, demographic, and infrastructural conditions of the Unified Health System. AI should ultimately be assessed not only by technical performance, but by its capacity to strengthen safe, equitable, evidence-based, and person-centered nursing care.