Sarah Wu

How do people make sense of the complex social world they are embedded in? To understand other people’s actions, two core cognitive mechanisms are necessary: counterfactual simulation and mental state inference. In my research, I develop a unified computational account of action explanation that draws on psychological insights and computational models of these two processes. I aim to leverage these ideas to enrich the capabilities of AI systems, and to better understand legal concepts like responsibility and reasonableness. Together, my research helps elucidate an important foundational component of human social cognition with meaningful applications to AI and the law.