To tackle increasingly complex tasks, AI must do more than compute—it needs to understand and anticipate human behavior. Be.FM, a new AI system developed in collaboration with U-M researchers, is the first designed specifically to predict, simulate, and reason about human actions. In this Michigan News research spotlight, Yutong Xie, a doctoral student in information science at U-M and the study’s lead author, explains how Be.FM is trained and shares their vision for its future applications.