Large language models (LLMs) in the evaluation of emergency radiology reports: performance of ChatGPT-4, Perplexity, and Bard

•LLMs, like GPTs, widely generate precise answers. •LLMs transform text reports for efficient communication. •Study assesses ChatGPT, Bard, Perplexity extracting emergency data, vs. human experts. •LLMs exhibit variable performance across medical specialties. •Study highlights algorithm, data, context, stresses ethical LLM integration in medicine. Re: Re: Large language models (LLMs) in evaluation of emergency radiology reports: performance of ChatGPT-4, Perplexity and Bard.Clinical RadiologyPreview Full-Text PDF