Problem-Solving

Abstract

Human and artificial agents are both intelligent problem solvers. Therefore, problem-solving will be central to their collaboration. Among notable developments in this domain is the capability for artificial agents to sample and search in a very farsighted fashion, or to be hyperopic, which is the technical term for farsighted vision, the opposite of myopia. This inverts the dominant concern of prior theory, which focuses on limited, bounded capabilities in problem-solving and decision-making. This shift poses significant opportunities and risks for augmented agents. Human processing will likely remain naturally myopic and limited, while artificial processing is increasingly hyperopic and powerful. Given these differences, digitally augmented problem-solving could be extremely divergent and dysfunctional, for example, by sampling and searching in a hyperopic fashion, while guided by persistent human myopia. Alternatively, one agent might dominate the other, leading to extreme convergence and possibly the digital domination of problem-solving.