Multidimensional manifestations of artificial intelligence literacy among clinical nurses: a qualitative descriptive study

Artificial intelligence (AI) is increasingly integrated into nursing practice, but its safe, critical, and responsible use depends on nurses’ AI literacy. Existing studies have mainly used quantitative approaches to assess overall levels of AI literacy, providing limited insight into how it is manifested in clinical practice. This study aimed to explore the multidimensional manifestations of AI literacy among clinical nurses from a qualitative perspective. This qualitative descriptive study was conducted in a tertiary Grade A general hospital in southwest China. Clinical nurses were recruited using maximum variation purposive sampling, considering age, years of experience, educational background, role, and department type. Data were collected from 26 February through March 2026 using semi-structured individual interviews and analysed using framework-based content analysis guided by four core AI literacy domains and an additional attitudinal domain. Reporting followed the Consolidated Criteria for Reporting Qualitative Research (COREQ). Twenty-three clinical nurses participated. Participants’ accounts suggested that AI literacy in clinical nursing was not merely a technical skill, but a practice-based capability shaped by real-task use, professional judgement, and institutional context. Nurses often approached AI through a “tool-first” logic: perceived usefulness in information retrieval, documentation support, patient education, and learning preceded systematic understanding of AI mechanisms. In practice, AI was mainly used as a human-led auxiliary resource; nurses valued its efficiency but emphasised the need to revise, verify, and contextualise AI-generated outputs. A prominent finding was ethical misalignment: some nurses had limited awareness of privacy and governance-related risks, yet strongly regarded themselves as responsible for the consequences of AI-assisted care. Attitudes towards AI were generally open but ambivalent, with expectations of workload relief and professional support coexisting with concerns about dependence, replacement, and disruption of clinician–patient trust. AI literacy among clinical nurses is multidimensional, context-embedded, and heterogeneous. It should be understood not simply as the ability to operate AI tools, but as a broader capability involving judgement, responsibility, and professional role negotiation. Future efforts should move beyond technical training alone to develop tailored, practice-oriented support strategies that emphasise safe and appropriate use, thereby facilitating effective AI integration into nursing practice.