Reflect-Compare-Revise (RCR): A Practitioner Inquiry on Integrating Generative AI in Nursing Care Planning Education

Generative AI can improve the linguistic fluency and organization of student writing, but dependence on it for answers may weaken clinical reasoning, raising concerns about AI-shaped nursing judgment and accountability. This practitioner action research developed and implemented a Reflect-Compare-Revise (RCR) approach for integrating ChatGPT into a nursing care plan (NCP) task using a standardized term-newborn case. Second-year nursing students (n = 40) produced three linked artifacts (human-only draft, ChatGPT output, and revised submission) and completed reflection journals; 15 students joined semi-structured interviews, and the teacher-researcher kept a reflexivity log. Artifacts were coded for accuracy, completeness, clinical appropriateness, verification actions, and revision depth, and reflections were thematically analyzed. RCR preserved cognitive ownership through human-first drafting, positioned AI output as a comparator, and prompted discrepancy-driven verification and revision. Revised NCPs more often showed cue-aligned diagnoses, newborn-appropriate targets and monitoring intervals, feasible shift-bounded actions, and auditable documentation checkpoints. Students valued organizational support but identified risks of generic output, overreliance, and confident misinformation, reinforcing the need for clear boundaries, privacy safeguards, and equitable access. Findings were translated into a classroom framework and policy guidelines that operationalize responsible GenAI use through mandatory human-first work, verification routines, transparent documentation, and assessment anchored on revision depth.

The online version contains supplementary material available at https://doi.org/10.1007/s40670-026-02736-1.