Numerical simulations of AI–human interaction using quintic mean-field Ising model

Abstract

The growing role of artificial intelligence (AI) in our society presents both benefits and challenges. To delve deeper into this, we suggest a simulation model that mimics a human–AI ecosystem. This model considers various factors, such as biased actions and multiple-agent interactions ranging from two to five agents. For instance, it includes situations where four humans engage with AI or where four AI systems interact with a human. Our main goal is to investigate how the number of AI systems in the ecosystem influences its dynamics. Our findings revealed that even slight variations in the presence of AI can cause significant changes in the system’s behavior. These shifts can push the system to distinct opposing outcomes or leave it in a state of uncertainty.