Modeling the Spread of Rumors in Social Networks Using LLMs
A recent challenge problem is to investigate and develop solutions that integrate LLM-driven prompts for agent-based models (ABMs). Using the NetLogo agent-based modeling platform we study whether simulations of the spread of rumors can be enhanced by LLM-based agents that decide on the salience of a rumor shared by its neighbors and chooses whether to further share the rumor to its neighbors. We investigate how LLM agents can enhance the study of alternative theories for the spread of mis- and dis-information in social networks and the ways to seed the prompts for such agents to generate different sharing behaviors.