I study how artificial intelligence changes financial information, who produces it, how it is measured, and how it shapes economic decisions. My research combines field surveys, large-scale behavioral and textual data, and computational methods to examine belief formation and decision-making by households, investors, workers, firms, and financial intermediaries.

My current work focuses on AI as both a source of financial information and a tool for interpreting it. One project combines an independently fielded expectations survey, browser-side behavioral data, and benchmarks from multiple large language models to examine how generative AI enters surveys as a respondent, an assistant to human respondents, and an analyst of the resulting data. I show that reliance on common models can create artificial consensus: reported expectations remain plausible at the center while disagreement and economically important tail signals are compressed.

This agenda builds on my earlier research in corporate and household finance, including how firms respond to transition opportunities and whether financial advisers align their recommendations with investors' preferences. Across these projects, I aim to develop rigorous and scalable measures of otherwise difficult-to-observe beliefs and behavior, and to clarify when new technologies improve financial decision-making—and when they instead distort the information on which those decisions rely.