Thomas Renault (RITM-Université Paris-Saclay) will present his paper “The Impact of Large Language Models on Research in Economics”.
Abstract: Large language models (LLMs) are rapidly transforming academic research, yet their adoption and effects on scholarly output in economics remain largely unexplored. Using comprehensive bibliometric data from OpenAlex covering millions of economics articles and working papers, we develop novel measures of LLM adoption combining a machine-learning classifier trained on synthetic LLM rewritings, linguistic markers in abstracts, and text-complexity measures. We document a sharp increase in LLM use and substantial heterogeneity in adoption across researchers, fields, institutions, and countries. Leveraging three complementary identification strategies – comparisons between adopters and non-adopters, differential incentives to adopt proxied by pre-LLM abstract complexity, and the temporary 2023 ChatGPT ban in Italy as a quasi-natural experiment – we estimate the causal effects of LLM use on both the volume and the quality of research output.
Link to the RITM Economics seminar web page