Visualizations for human-machine teams in complex environments: design concepts and review of current approaches

Abstract

Recent advances in artificial intelligence expand the possibilities for partnering humans and machines in complex environments. However, the design of visualizations on interfaces for supporting effective communication between humans and machines remains a significant challenge. In this paper, we explore the problem of designing visualizations to support effective human-machine interaction especially in complex environments with high levels of instability and ambiguity. We review current approaches to visualization design and articulate the limitations of these approaches for our problem space. We propose an approach that focuses the design of visualizations on system constraints, or event-independent boundaries of successful operation, rather than tasks and user characteristics. This approach goes beyond specific events and individuals to cover a larger space of possibilities for action by human-machine teams in complex environments. These concepts can be used to design visualizations that provide value to human-machine teams even in the face of uncertainty. The results of this work provide a starting point for future research addressing shortfalls in human-machine interactions which can have real-world consequences.