Man versus Machine or Man + Machine?

Allocating roles and functions between the human and computer is critical in defining efficient and effective system architectures. However, past methodologies for balancing the roles and functionalities between humans and computers in complex systems have little connection to different types of required cognition, behaviors, or tasks, or don´t address the role of uncertainty in the environment. To augment these previous role allocation approaches, this article presents a modification to the skill, rule, and knowledge-based behavior taxonomy that includes expertise and uncertainty. Skill-based behaviors are the best candidates for automation, assuming significant sensor performance assumptions can be met, but rule and knowledge-based reasoning are better suited for human-computer collaboration. Such systems should be designed so that humans harness the raw computational and search power of computers for state-space reduction, but also allow them the latitude to apply their expertise in uncertain situations through inductive reasoning for potentially creative, out-of-the-box thinking.