MACOMB, IL - - Western Illinois University Cecil P. McDonough Endowed Business Professor J. Jobu Babin was recently invited to present his research to a select group of academics and policy-makers at the Federal Reserve Bank of New York.
His research in collaboration with WIU Professor Haritima Chauhan, "Chatbot or Humanaut? How the Source of Advice Impacts Behavior in One-shot Social Dilemmas,’” explores the impact of types of advice in three classic scenarios in behavioral game theory, employing machine learning tools from economics, computer science and information science. The authors find that people in these social dilemmas prefer (and are more willing to pay for) naive human advice rather than ChatGPT4-generated advice. This preference results in increased prosocial behavior and greater social efficiency – a finding with significant implications for how institutions use AI.
“Our results are critical in an age in which people increasingly solicit advice from AI in situations, ranging from portfolio management to ethical dilemmas, where humans were previously the only option,” Babin said. “The research may help shed light on markets in which chatbots are costing humans jobs and situations where humans can create more socially efficient outcomes.”