When the answer is right, who did the reasoning? Rethinking clinical judgment in the age of generative AI.

Abstract

Generative artificial intelligence can now produce plausible and sometimes correct responses to complex learning tasks, making answer quality a less reliable indicator of students' understanding and judgment. When artificial intelligence contributes to information seeking, synthesis, interpretation, and answer generation, the final product may not show how much of the underlying reasoning can be attributed to the learner. This is particularly important in nursing, where students will ultimately be accountable for clinical decisions. The educational challenge is therefore not only to promote critical use of artificial intelligence, but also to establish credible evidence that clinical judgment belongs to the learner. Teaching should require students to question, verify, contextualize, and justify information generated by artificial intelligence, while assessment should make reasoning observable rather than infer competence from the final answer alone. Two contrasting examples illustrate this approach. In type 1 diabetes care, students reconstruct and defend the reasoning behind quantitative clinical decisions; in dementia care, they demonstrate relational judgment through real-time adaptation. As answer generation becomes increasingly automated, clinical judgment should be judged by what students can independently explain, adapt, and defend.