JMIR Publications

ABSTRACT

Background:

Generative artificial intelligence (AI) is increasingly shaping how patients understand health information before clinical encounters, creating challenges for patient-professional communication involving professional authority, responsibility, uncertainty, and trust. However, little is known about how nursing students interpret these issues as they move from conceptual preparation into clinical practice.

Objective:

This study aimed to examine how nursing students’ interpretations of ethical issues in AI-mediated clinical communication were maintained, challenged, or reconsidered across classroom engagement, clinical placement, and subsequent reflection.

Methods:

We conducted a prospective longitudinal mixed methods study involving 32 third-year undergraduate nursing students, including 16 international and 16 local Chinese students enrolled in the same Bachelor of Nursing program. Questionnaire data were integrated with post-session written responses, clinical reflective logs, and post-placement group reflections across sequential classroom, clinical, and reflective contexts. Qualitative evidence constituted the primary interpretive strand, while questionnaire findings provided complementary contextual evidence. Participant-level evidence was linked longitudinally and analyzed using reflexive thematic analysis, within-case interpretation, and cross-case comparison.

Results:

Among the 32 participants, clinical experience complicated rather than simply extended interpretations expressed after classroom engagement. Participants encountered tensions involving the negotiation of professional and AI-informed interpretations, professional responsibilities beyond factual correction, and the communication of uncertainty while maintaining patient trust. Participant-linked longitudinal analysis identified recurring patterns of broad reconsideration across ethical issues, selective reconsideration in particular areas, and persistent or unresolved tensions. Clinical experience and subsequent reflection therefore sometimes made competing ethical considerations more visible without producing greater certainty or uniform change.

Conclusions:

AI-mediated clinical communication presents challenges extending beyond the accuracy of AI-generated health information to the ways in which AI-informed patient understandings interact with professional authority, responsibility, uncertainty, and trust. Preparing nurses for AI-mediated clinical communication may therefore require attention not only to the evaluation of AI-generated information but also to the relational and ethical tensions that arise when such information enters clinical encounters.