Seminar "Artificially Intelligent Economic Agents: Two Cases from the History of the Santa Fe Institute"

Murat Bakeev (Leibniz University Hannover) presented a paper titled "Artificially Intelligent Economic Agents: Two Cases from the History of the Santa Fe Institute" on June 11, 2025.

Abstract:

This paper reconstructs two attempts to transfer classifier systems from computer science to economics, with the aim of using them in modeling economic agents. Both initiatives took place under the auspices of the Santa Fe Institute’s Economics Program, which ran from 1988 to 2004. Classifier systems, developed by computer scientist John Holland, are a reinforcement learning model whose mechanism is based on the evolutionary selection and modification of competing rules that determine the actions of agents and the conditions under which they are applied. The first case examined is a model of the emergence of money as a medium of exchange, constructed by Ramon Marimon, Ellen McGrattan, and Thomas J. Sargent. The second is the Santa Fe Artificial Stock Market, created by W. Brian Arthur, John H. Holland, Blake LeBaron, Richard G. Palmer, and Paul Tayler. Using archival material from the Kenneth Arrow Collection and oral interviews, the paper analyses the motivation behind these interdisciplinary transfers and the reasons that have prevented the further spread of Holland's classifier systems in economic modelling.