Artificial intelligence-mediated cyberincivility in health professions education: An evolutionary concept analysis

The pervasive embedding of generative artificial intelligence into health professions education has introduced novel forms of harm that challenge how digital professionalism is taught and evaluated. Traditional definitions of cyberincivility emphasize human intent and relational aggression; however, generative artificial intelligence systems lack consciousness yet can produce biased or inaccurate outputs that affect technology-mediated learning and clinical training.

To (a) examine how cyberincivility is evolving in artificial intelligence-enabled health professions education contexts, and (b) define and clarify artificial intelligence-mediated cyberincivility to support the ethical adoption of artificial intelligence, including related curriculum development and institutional governance.

Rodgers' evolutionary concept analysis.

MEDLINE (via PubMed), CINAHL, Scopus, Web of Science Core Collection, PsycINFO, ERIC, ProQuest Dissertations and Theses Global, and IEEE Xplore were searched from January 2015 to November 2025.

Data on uses of the concept, surrogate and related terms, attributes, antecedents, consequences, and contextual variations were extracted and synthesized.

Thirty-two studies published between 2021 and 2025 met inclusion criteria. Artificial intelligence-mediated cyberincivility was rarely named explicitly and was most often described through the mechanisms that produce it, including algorithmic bias, hallucination, and false information. Two attributes were supported across a substantial proportion of the corpus: representational bias and stereotyping, and epistemic incivility. Three further attributes, symbolic exclusion, tone and relational incivility, and evaluative harm, were identified in few sources and are reported as provisional. Antecedents were organized across four levels: structural, systemic, epistemic, and interactive. Consequences extended across learner, educator, patient-facing, and system or institutional levels, with learner-level consequences most frequently reported. Artificial intelligence-mediated cyberincivility in health professions education is proposed as a phenomenon of impact, independent of intent, in which outputs and evaluative interactions appear professionally acceptable while producing biased or differential content, plausible but incorrect information delivered in a register that suppresses verification, representational skew, or diminished relational depth that may be experienced as dehumanizing.

This analysis proposes extending the conceptual boundaries of cyberincivility beyond human intent, and describes how generative systems may mislead users or produce interactions experienced as dehumanizing, often through polished and affirming language. The concept is provisional and derived from a heterogeneous and emerging evidence base. Preserving professional integrity in artificial intelligence-enabled health professions education is likely to require an evolution of digital professionalism that includes the capacity to interrogate artificial intelligence outputs, strengthen learner epistemic agency, and keep automated evaluation open to question and appeal.

Artificial intelligence systems do not have "intent" but may still cause "incivility." From hallucinations to algorithmic bias, our evolutionary concept analysis maps how artificial intelligence may harm learners and patients through polished, professional-sounding communication. Digital professionalism needs to evolve with it.