Living in interesting times: Some opportunities and threats from AI in quantitative social sciences

Large Language Models (LLMs) are playing an increasingly large role in quantitative social science research. This includes idea generation, writing code, acting as study participants, and even writing and reviewing papers. In this talk Rohan Alexander will discuss some of the threats and opportunities this change brings. He will begin by discussing their role in improving reproducibility and then discuss some limitations. He will then discuss some of his current research on generating and evaluating quantitative social science with LLMs. He will conclude by highlighting some key elements needed to do research with AI in the quantitative social sciences.

Rohan Alexander is an Assistant Professor at the University of Toronto, jointly appointed in the Faculty of Information and the Department of Statistical Sciences. He is interested in developing workflows that improve the trustworthiness of data science and tends to be especially focused on the role of code and testing. He holds a PhD in Economic History from the Australian National University and is the author of the book Telling Stories with Data.