Advancement to Candidacy Presentation: “The Who, What, When, and Why of AI Use in Online Experiments”, Austin Brooksby, University of California, Santa Barbara
Speaker
Austin Brooksby, PhD Student, University of California, Santa Barbara
Biography
Austin holds a B.S. in Economics from Utah State University, an M.S. in Finance from the University of Utah, and an M.A. in Economics from the University of California, Santa Barbara. His academic and research background spans experimental and behavioral economics, market design, game theory, and algorithmic trading behavior.
As an undergraduate researcher at Utah State University, Austin worked with Dr. James Feigenbaum on behavioral macroeconomic modeling and wealth taxation, contributing to published policy papers on immigration and wealth inequality. Also during his time at Utah State University, he authored work on experimental tests of a novel social preference theory with Dr. Lucas Rentschler, Dr. Vernon Smith, Jacob Meyer, and Robbie Spofford. He also served as an Undergraduate Research Fellow at the Center for Growth and Opportunity, where he applied econometric methods to studies on electricity policy and innovation.
He later joined the University of Utah’s Lab of Experimental Economics and Finance as a Research Assistant working with Dr. Elena Asparouhova, administering market simulations, analyzing data with Python and R, and supporting experimental design. In addition, he has experience teaching graduate finance courses as a Teaching Assistant,reinforcing his dedication to both research and education in economics and finance.
Austin employs behavioral theory and experiments to study how bounded rationality aggregates from individual to group behavior, and the beliefs that individuals hold over the behavior of groups in the presence of institutions such as markets. Another vein of his research uses experiments to study people's underlying models- or causal understandings of the complex world around them- and how these models inform their support for economically important environmental policies. Finally, he also studies human/AI interaction, specifically the determinants of AI-seeking behavior in economics experiments, and what information people seek from AI assistants.
Title
“The Who, What, When, and Why of AI Use in Online Experiments”
Abstract
Online experiments are now a core empirical tool in economics, but ubiquitous AI assistants raise a practical and conceptual challenge: subjects may consult AI while completing tasks intended to elicit preferences. We study who uses AI in preference experiments, what they use it for, when they use it, and how AI availability affects elicited preferences. We conduct three canonical preference elicitations- risk, time, and social preferences- using a standardized workflow and a within-subject introduction of an embedded AI chatbot assistant. Our central interpretation is conceptual: because the experiment already provides payoff-relevant information about the choice objects, AI use inside these tasks cannot be motivated by information acquisition about the objects per se. Therefore, systematic within-subject changes in choices following AI availability are interpreted as evidence of revealed incompleteness of preferences in the sense of Nielsen and Rigotti, 2026. A primary goal is comparative: we test whether AI take-up and AI-induced preference changes differ systematically across domains, consistent with the hypothesis that the preferences elicited with our standard experimental methods are less complete in some domains than others.
JEL Codes: C91, D91
Event Details
Austin will be presenting his Advancement to Candidacy paper, “The Who, What, When, and Why of AI Use in Online Experiments”. To access the Advancement paper, you must have an active UCSB NetID and password.
PLEASE DO NOT CIRCULATE, PRELIMINARY WORK
