Medical students report coherent ethical orientations toward ai across academic and clinical domains, with no clear association with perceived technical readiness: a cross-sectional pilot study

Abstract Background Whether perceived technical AI competency is associated with greater ethical openness to AI integration, a relationship often assumed in curricular design, remains empirically untested. This pilot study examined relationships between perceived AI readiness, ethical perspectives toward AI integration, and curricular needs among medical students. Methods A cross-sectional survey was administered to medical students at a single U.S. institution from September through November 2025. The survey assessed self-reported AI readiness, ethical perspectives across academic and clinical domains, and responses to 22 scenario-based ethical items; all measures were student self-reports, and findings reflect student perceptions. Composite indices were created for Perceived AI Readiness, AI Integration Receptivity, Academic Ethics, and Clinical Ethics. Spearman correlations and Wilcoxon rank-sum tests were used, with Bonferroni correction for 18 tests (α = 0.0028). Responses to three open-ended items were summarized descriptively. Results Among 45 respondents (7.5% response rate), perceived AI readiness was not significantly associated with AI integration receptivity (ρ = 0.118, 95% CI [− 0.191, 0.410], p = 0.4419); no clear association between perceived technical competency and ethical receptivity was detected in this pilot sample. However, academic and clinical ethics perspectives were strongly correlated (ρ = 0.719, p < 0.0001), consistent with a coherent ethical orientation across domains. No clear training-phase difference in ethical perspectives was detected, although the small clinical subgroup limited precision. Scenario analysis revealed unanimous consensus that direct cheating and unauthorized patient data exposure are unethical, while scenarios involving AI-human collaboration and de-identification practices generated disagreement. Open-ended responses (30 of 45 respondents) showed that supporters and opponents of AI integration alike requested instruction on AI hallucination and limitations, patient data privacy rules, and explicit assessment boundaries. Conclusions Participating medical students reported nuanced ethical reasoning about AI integration, distinguishing clear violations from contested applications. The absence of a detectable association between perceived AI readiness and AI integration receptivity provides no support for the assumption that technical familiarity promotes ethical acceptance, although the limited reliability of the AI Readiness Index and the small sample preclude firm conclusions. The coherence of ethical orientations across academic and clinical contexts is a hypothesis-generating association that warrants longitudinal investigation.