A Causal Model of Theory of Mind in Conflict for Artificial Intelligence

This paper introduces a structural causal model of Theory of Mind (ToM) in conflict situations, formalized as a directed acyclic graph (DAG). It identifies four exogenous variables and five endogenous mediators that influence ToM engagement, leading to improved epistemic accuracy through three causal pathways: tractability, reasoning depth, and enabling causes. This framework allows AI systems to engage in mentalizing in a resource-efficient manner, which is crucial for enhancing human-machine collaboration and developing robust social intelligence in AI applications.