WHAT’S IN THE CHATTERBOX? LARGE LANGUAGE MODELS, WHY THEY MATTER, AND WHAT WE SHOULD DO ABOUT THEM

Large language models (LLMs)—machine learning algorithms that can recognize, summarize, translate, predict, and generate human languages on the basis of very large text-based datasets—are likely to provide the most convincing computer-generated imitation of human language yet. Because language generated by LLMs will be more sophisticated and human-like than their predecessors, and because they perform better on tasks for which they have not been explicitly trained, we expect that they will be widely used. Policymakers might use them to assess public sentiment about pending legislation, patients could summarize and evaluate the state of biomedical knowledge to empower their interactions with healthcare professionals, and scientists could translate research findings across languages. In sum, LLMs have the potential to transform how and with whom we communicate.The Technology Assessment Project is supported in part through a generous grant from the Alfred P. Sloan Foundation (grant #G-2021-16769)http://deepblue.lib.umich.edu/bitstream/2027.42/191718/1/large-language-models-TAP-2022-final-051622.pdfhttp://deepblue.lib.umich.edu/bitstream/2027.42/191718/2/LLMImplicationsforScience.pdfhttp://deepblue.lib.umich.edu/bitstream/2027.42/191718/3/Large Language Models Executive Summary 2022.pdfhttp://deepblue.lib.umich.edu/bitstream/2027.42/191718/4/large-language-models-one-pager STPP-TAP-2022-v3.pdfDescription of large-language-models-TAP-2022-final-051622.pdf : What’s in the Chatterbox? Large Language Models, Why They Matter, and What We Should Do About ThemDescription of LLMImplicationsforScience.pdf : Implications for the Scientific Landscape (31 pages)Description of Large Language Models Executive Summary 2022.pdf : Executive Summary- LLMDescription of large-language-models-one-pager STPP-TAP-2022-v3.pdf : One-pager LLMSEL