Studies and Analyses of Aided Adversarial Decision Making. Phase 2: Research on Human Trust in Automation

摘要

This report describes the second phase of work conducted at the Center for Multi-source Information Fusion at the State University of New York at Buffalo. This work focused on Aided Adversarial Decision Making (AADM) in Information Warfare (IW) environments. Previous work examined informational dependencies and vulnerabilities in AADM to offensive IW operations. In particular, human trust in automated, information warfare environments was identified as a factor which may contribute to these vulnerabilities and dependencies. Given that offensive IW operations may interfere with automated, data-fusion based decision aids, it is necessary to understand how personnel may rely on or trust these aids when appropriate (e.g., when the information provided by the aids is sound), and recognize the need to seek other information (i.e., to 'distrust' the aid) when the information system has been attacked. To address these questions, this report details background research in the areas of human trust in automated systems and sociological findings on human trust, details the development of an empirically-based scale to measure trust, provides a framework for investigating issues of human trust and its effect on performance in an AADM-IW environment, and describes the requirements for a laboratory designed to conduct these investigations.