Title: Models, Metrics, and a Mathematics of Interactional Trust for Humans and Automation

Yosef Razin

Robotics Ph.D. Student

School of Aerospace Engineering

College of Engineering

Georgia Institute of Technology

Date: October 30, 2018 (Tuesday)

Time: 2-4pm EST

Location: MK 317

Committee:

Dr. Karen Feigh, School of Aerospace Engineering, Georgia Tech

Dr. Ayana Howard, School of Interactive Computing, Georgia Tech

Dr. Jason Borenstein, School of Public Policy, Georgia Tech

Dr. David Gefen, LeBow College of Business, Drexel University

Dr. John D Lee, Industrial and Systems Engineering, University of Wisconsin-Madison

Abstract: 

Trust, the glue of society and foundation of morality, according to both ancient sages and modern scholars, despite its power, has proven elusive to the academy. Difficult to capture in a model or pin down with metrics, researchers have tracked trust in over two dozen fields and cited it in tens of thousands of publications. While some consensus has emerged on its major attributes and some fields have even coalesced around singular definitions, many remain with a fragmentary picture of trust. Nowhere is this truer than in human-automation interaction, which concluded that trust, elsewhere a force of robustness, is actually fragile. This proposal aims to resolve this paradox by reframing trust in human-automation interaction, contextualizing it within the larger academic dialogue. A model of interactional trust will be developed that has greater explanatory power and is more consistent with other fields. Metrics of fragility and trust beliefs and decisions will be presented and validated. A mathematics of trust, based in Subjective Logic and tying together a number of disparate trust theories, will be proffered and proven. Thus armed, three experiments in the driving and gaming domains will be carried out and used to refine, demonstrate, and validate the proposed theoretical contributions.