2023 WMIF | Disruptive Dozen #5

Harnessing the power of large language models to improve health care

Large language models (LLMs) are a form of artificial intelligence that work by sifting through massive datasets to discern patterns and relationships among words — a process known as training. Once these models are sufficiently trained, they can perform a variety of language-based functions, including recognizing, summarizing, translating, generating, and predicting text. LLMs could have a significant impact in medicine by streamlining and supporting the work of physicians. Two key areas include clinical decision support and administrative workflows — offering information and suggestions at the point of care and reducing the time spent on manual, repetitive tasks.