Ashesh Rambachan announces a new working paper co-authored with Sendhil Mullainathan and Jens Ludwig that explores the use of large language model (LLM) outputs in empirical research. The post signals a forthcoming thread with more details and links to the full paper.

This work is notable given the growing interest in integrating LLMs into social science and economics research workflows. The collaboration brings together prominent researchers in economics and machine learning, suggesting the paper likely addresses methodological considerations and best practices for leveraging AI-generated outputs in rigorous empirical analysis.