Indirect Reciprocity in Hybrid Human-AI Populations

Indirect reciprocity (IR) is a key mechanism to explain cooperation in human populations. With IR, individuals are associated with reputations, which can be used by others when deciding to cooperate or defect: the costs of cooperation can therefore be outweighed by the long-term benefits of keeping a specific reputation. IR has been studied assuming human populations. However, social interactions involve nowa- days artificial agents (AAs) such as social bots, conversational agents, or even collaborative robots. It remains unclear how IR dynamics will be affected once artificial agents co-exist with humans. In this project we aim to develop game-theoretical models to investigate the potential effect of AAs in the dynamics of human cooperation. We study settings where artificial agents are potentially subject to the different reputation update rules as the remaining individuals in the population. Furthermore, we consider both settings where reputations are public and setting where reputations are privately held. We show that introducing a small fraction of AAs, with a strategy discriminating based on reputation, increases the cooperation rate in the whole population. Our theoretical work contributes to identify the settings where artificial agents, even with simple hard-coded strategies, can help humans solve social dilemmas of cooperation. At this workshop, we hope to discuss future research avenues where citizens preferences, incentives, and strategic adaptation are considered when designing artificial agents to leverage cooperation in hybrid systems.