How Does Interaction with Artificial Agents Affect Prosociality among Humans?

• In Event: Theory and Experimental Evidence of Cooperation and Coordination

Abstract

Increasingly, humans are interacting with artificial agents. There is a long history of employing computerized agents in experimental economic games, mostly for the purposes of experimental control (See March, 2021 in the Journal of Economic Psychology for a comprehensive review). Recently social scientists have begun to directly study the effects of interacting with artificial agents on cooperative and prosocial behavior. These research designs typically involve comparing the behavior when playing with an artificial agent to behavior when playing against a human. Previous research has shown that humans are more selfish when interacting with artificial agents than when interacting with humans. This previous research typically looks at one-shot interactions. Here we are interested in the downstream effects on human behavior from interacting with an artificial agent. Cooperative and prosocial behavior can be influenced by the actions taken by others in previous interactions—in other words, cooperation begets further cooperation. The result is the emergence of prosocial behavior cascades. We will use trust, ultimatum, and public goods games to test whether behavior in a network that includes both artificial agents and humans can lead to contagions in the same way that purely human networks have been shown to do.

Author

• Reuben Kline, SUNY, Stony Brook University