Change in Human (Moral) Decision-Making and Performance with Intelligent Machines: The Good, the Bad, and the Ugly in Human-Autonomous Systems Interactions

Abstract

Autonomous systems and intelligent machines are involved in almost all areas of human activity and they are now more and more present in our everyday life. The reason for this extensive use certainly resides in all the benefits these machines offer to the users. In experimental settings, numerous studies have demonstrated the positive effects that the introduction of autonomous systems have on human decision-making and performance. However, studies have shown in addition that the introduction of these systems can have important negative effects as well. Considering that autonomous systems are now introduced in sensitive domains like the military or medicine, we need more than ever a comprehensive understanding of the effects they cause on human performance and decision-making, and particularly in tasks and contexts with a social or moral dimension. The aim of this narrative review is threefold. First, we will provide an overview of the main effects on a human agent’s decision-making and performance produced by the introduction of autonomous systems. Second, we will review the conditions identified as underlying factors of these effects, and see how current models of human – autonomous systems interaction integrate those conditions. Third, we will conclude this review by highlighting new directions for future investigations.