What’s the Difference? Reconciling Knowledge Structure Concepts to Aid AI Development for Human-Machine Teaming

Abstract

A significant challenge for the development of artificial social intelligence for effective human-machine teams is defining the forms of artificial knowledge structures needed for machine agents to meaningfully engage in collaboration. Relevant to this, individual and shared knowledge structure concepts have been proposed across a variety of disciplines, resulting in a lack of conceptual clarity and impeding their operationalization for human-machine teaming. To reconcile conceptual differences across disciplines and enable the emergence of complex socio-cognitive abilities in machine agents, research is needed to integrate theory on the knowledge structures that underpin complex cognition. Toward this end, we survey research from the cognitive and computational sciences to develop a framework for the systematic application and evaluation of knowledge structure concepts for machine agents in teams. Our approach focuses on contextual factors, specifically the task environment structure and the situation temporality, that can help guide knowledge structure requirements for artificial social intelligence.