The American Journal of Emergency Medicine's policy on large language model usage in manuscript preparation: Balancing innovation and responsibility

Abstract

Large Language Models (LLMs) represent a transformative advancement in the preparation of medical scientific manuscripts, offering significant benefits such as reducing drafting time, enhancing linguistic precision, and aiding non-native English speakers. These models, which generate text by learning from extensive datasets, can streamline the publication process and maintain consistency across collaborative projects. However, their limitations, including the risk of generating plausible yet incorrect text and the potential for biases, necessitate careful oversight. Ethical concerns about accuracy, authorship, and transparency need to be carefully considered. The American Journal of Emergency Medicine has adopted a policy permitting LLM use with full disclosure and author responsibility, emphasizing the need for ongoing policy evolution in response to technological advancements.