AI 中文总结
本研究开发智能体式检索增强型LLM CORA，发现其可将46名医师的诊断准确率从70.8%提升至82.6%，但存在错误答案获引用支持时医师抵抗大幅下降的安全风险。
AI 中文摘要
检索增强型大型语言模型（LLMs）可提供关联来源的临床支持，但其价值取决于展示的证据是引导还是扭曲医师的依赖。我们开发了CORA，一种智能体式检索增强型LLM，以研究关联来源的辅助如何影响医师决策。CORA保持了基准性能，且在模型训练数据截止后发布的病例上取得了更大提升。在一项包含46名医师的研究中，无辅助时准确率为70.8%，使用CORA后升至82.6%。支持性引用可预测正确答案（87.7%对65.5%），但引用造成了重要的不对称性：感知到的支持使医师采纳正确建议的比例从34%升至76.9%，但当LLM的错误答案有引用支持时，医师对其的抵抗从92%降至34.8%。这些发现表明，关联来源的LLM辅助可提高医师准确率，同时引入了依赖于依据的安全风险。
英文摘要
Retrieval-augmented large language models (LLMs) promise source-linked clinical support, but their value depends on whether displayed evidence guides rather than distorts physician reliance. We developed CORA, an agentic retrieval-augmented LLM, to investigate how source-linked assistance affects physician decision-making. CORA maintained benchmark performance and achieved larger gains on cases published after the models' training-data cutoffs. In a study of 46 physicians, accuracy increased from 70.8% unaided to 82.6% with CORA. Supporting citations predicted correct answers (87.7% vs 65.5%), but citations created an important asymmetry: perceived support increased adoption of correct advice from 34% to 76.9% but when an incorrect LLM answer appeared citation-supported, physician resistance to it fell from 92% to 34.8%. These findings show that source-linked LLM assistance can improve physician accuracy while introducing a grounding-dependent safety risk.
