Establishing Causality in Complex Human Interactions: Identifying Breakdowns of Intentionality

People in complex scenarios face the challenge of understanding the purpose and effect of other human and computational behaviour on their own goals through intent recognition. They are left asking what caused person or system ‘x’ to do that? The necessity to provide this support human-computer interaction has increased alongside the deployment of autonomous systems that are to some degree unsupervised. This paper aims to examine intent recognition as a form of decision making about causality in complex systems. By finding the needs and limitations of this decision mechanism it is hoped this can be applied to the design of systems to support the awareness of information cues and reduce the number of intent recognition breakdowns between people and autonomous systems. The paper outlines theoretical foundations for this approach using simulation theory and process models of intention. The notion of breakdowns is then applied to intent recognition breakdowns in a diary study to gain insight into the phenomena.