医学中大型语言模型的环境影响。

Environmental impact of large language models in medicine.

作者信息

Kleinig Oliver, Sinhal Shreyans, Khurram Rushan, Gao Christina, Spajic Luke, Zannettino Andrew, Schnitzler Margaret, Guo Christina, Zaman Sarah, Smallbone Harry, Ittimani Mana, Chan Weng Onn, Stretton Brandon, Godber Harry, Chan Justin, Turner Richard C, Warren Leigh R, Clarke Jonathan, Sivagangabalan Gopal, Marshall-Webb Matthew, Moseley Genevieve, Driscoll Simon, Kovoor Pramesh, Chow Clara K, Luo Yuchen, Thiagalingam Aravinda, Zaka Ammar, Gould Paul, Ramponi Fabio, Gupta Aashray, Kovoor Joshua G, Bacchi Stephen

机构信息

School of Medicine, University of Adelaide, Adelaide, South Australia, Australia.

Division of Medicine, Lyell McEwin Hospital, Adelaide, South Australia, Australia.

出版信息

Intern Med J. 2024 Dec;54(12):2083-2086. doi: 10.1111/imj.16549. Epub 2024 Nov 14.

DOI:10.1111/imj.16549

PMID:39542015

Abstract

The environmental impact of large language models (LLMs) in medicine spans carbon emission, water consumption and rare mineral usage. Prior-generation LLMs, such as GPT-3, already have concerning environmental impacts. Next-generation LLMs, such as GPT-4, are more energy intensive and used frequently, posing potentially significant environmental harms. We propose a five-step pathway for clinical researchers to minimise the environmental impact of the natural language algorithms they create.

摘要

大语言模型在医学领域的环境影响涉及碳排放、水资源消耗和稀有矿物质使用。诸如GPT-3等前代大语言模型已经产生了令人担忧的环境影响。诸如GPT-4等下一代大语言模型能源消耗更高且使用频繁，可能造成重大环境危害。我们为临床研究人员提出了一个五步路径，以尽量减少他们所创建的自然语言算法对环境的影响。