Chatbot or Humanaut? How the Source of Advice Impacts Behavior in One-shot Social Dilemmas

Abstract This paper explores how the source of advice – human or generative AI (genAI) – impacts behavior in three classic social dilemmas. Utilizing a novel experiment, we show that, when it comes to prosociality, the source of advice matters. Players preferred human advice over that from genAI and were more willing to pay for it. Prosocial behavior was more prevalent when players received human advice, increasing cooperation in stag hunts and driving contributions in public goods and dictator games. Potential explanations include genAI’s empathy limitations and potential moral wiggle room. Leveraging natural language processing advances through large language models topic modeling, we demonstrate that the advice corpora differ significantly. Human advice was more objective, specific, intuitive, and norm-oriented; genAI offered reasoning and targeted concepts of risk. We propose that the adoption of genAI technologies should include an investment in human connectivity and organizational input with a goal of the constrained agency.