Adopting the Theory of Distributed Cognition for Human-AI Cooperation

Abstract

Questions around human-AI cooperation are increasingly important as AI begins to act as a teammate as opposed to as a tool used by human operators. To effectively design for this, a deeper understanding of human teams is required. Distributed Cognition (DC) is a theoretical framework used to study how groups of humans cooperate. It views cognition as not confined solely to an individual's mind but distributed across individuals, artifacts, and tools in the environment. This short paper applies DC to the study of human-AI cooperation by analysing cooperative practices in two boardgames of particular interest to the AI community. Using DC we show how humans utilise the physical environment to ..