Man-Machine teaming: a problem of Agency

Automation technology is a ubiquitous phenomenon that is profoundly changing our everyday life. While the public is generally enthusiastic about the possibilities that it offers, recent tragedies remind us of the difficulties human operators have in cooperating with highly automated systems. This issue of cooperation amongst team (and/with automates) has led research on (team) performance and situation awareness investigating how computer might support collaboration between operators. However, after decades of research, the cognitive coupling between human and machine remains difficult to achieve. In this paper, we outline that the recent explosion of interest in the experience of being an agent (agency) opens interesting novel avenues to explain and compensate such difficulties. The sense of agency refers to the feeling of control over actions and their consequences. In the first part of this paper, we present some works indicating that automation technology can alter the agentive experience dramatically. Then, we discuss why such change in agency can dramatically impact operator performance and system acceptability. In the last part of this article, we propose to apply the framework of agency to the HMI domain and to take into account how the information provided by an automated system influences how an operator understand and control such system. Taken together, the different studies presented suggest that the science of agency provides us new conceptual tools and measures to analyze agent-system interaction. By using these tools, engineers could design more acceptable and more controllable automated interfaces and optimize human-automation cooperation. (C) 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.