Challenges of human—machine collaboration in risky decision-making

Abstract

The purpose of this paper is to delineate the research challenges of human—machine collaboration in risky decision-making. Technological advances in machine intelligence have enabled a growing number of applications in human—machine collaborative decision-making. Therefore, it is desirable to achieve superior performance by fully leveraging human and machine capabilities. In risky decision-making, a human decision-maker is vulnerable to cognitive biases when judging the possible outcomes of a risky event, whereas a machine decision-maker cannot handle new and dynamic contexts with incomplete information well. We first summarize features of risky decision-making and possible biases of human decision-makers therein. Then, we argue the necessity and urgency of advancing human—machine collaboration in risky decision-making. Afterward, we review the literature on human—machine collaboration in a general decision context, from the perspectives of human—machine organization, relationship, and collaboration. Lastly, we propose challenges of enhancing human—machine communication and teamwork in risky decision-making, followed by future research avenues.