This project investigates how generative AI agents influence human behavior in social dilemmas, situations where individually rational decisions can lead to collectively undesirable outcomes. While existing research has largely focused on one-to-one human–AI interactions, this study examines more realistic settings in which people interact with and are influenced by multiple AI agents simultaneously. By exploring how AI-generated social information affects cooperation, decision-making, and social learning, the project seeks to understand how human behavior may evolve in increasingly AI-mediated environments. The findings will help guide the design of AI systems that participate in social settings as advisors, coordinators, or monitors, while preserving key human values such as agency, consent, and adaptability.