Answers are grounded in this paper and cite their sources.
Since the public release of ChatGPT in late 2022, large language models (LLMs) have transformed clinical medicine1. While advanced models now match or exceed trained physicians on nephrology board certification examinations2, a critical limitation of existing research is the reliance on multiple-choice questions (MCQs)2,3. Evidence that LLM access improves clinician diagnostic accuracy in realistic settings remains limited and mixed, with early randomized trials reporting divergent results4,5. In nephrology specifically, no randomized study had evaluated AI assistance using complex, multimodal, open-ended vignettes, leaving the clinical relevance of benchmark results uncertain.
No takes yet. Share an insight, caveat, or question.
Bentegeac et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: