3.0 The History (and Residue) of Measuring Trust in AI Episode

TRUST CONTROL: Trust is no longer a feeling.

3.0 The History (and Residue) of Measuring Trust in AI Episode

Roger C. Mayer and his fellow researchers introduced an influential trust model in 1995 in the Academy of Management Review. The impetus for their work? Corporate managers worried they were losing control of workers amid big changes. There were more people of color and women entering the workforce, plus workers upheaval in emerging self-directed teams could not be monitored as they once were. Little did Mayer and his co-authors James H. Davis and F. David Schoorman know that decades later their ABI (Ability, Benevolence and Integrity) trust model would become heavily influential in trust in AI related research. 



This episode of Trust Control is all about the historical ideas and "residue" that perpetuate today throughout trust measurement. Why do we see the same old trust measurement ideas, applied in questionable ways?

In this episode: Corporate motivations for measuring worker trust and trustworthiness Why the Mayer ABI model may not be right for AI How antiquated and sometimes poorly-tested ideas get baked into today's trust measurement Under- and over-trusting AI W- E- I -R -D history (Western Educated Industrialized Rich Democratic) of trust measurement Military funding of trust measurement from the Cold War to today

Featured in this episode: Kate Kaye, Trust Control creator and narrator Roger C. Mayer, Professor of Leadership in the Poole College of Management at North Carolina State University Bran Knowles, Professor of Sociotechnical Systems at Lancaster University, UK Gale Lucas, Director of the Technology Evaluation Lab at the University of Southern California's Institute for Creative Technologies, Research Associate Professor in USC's Departments of Civil Engineering, Computer Science and Psychology Meredith McNeil, Senior Research Scientist and Team Leader of the Genome to Phenome Team at Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia Andreas Duenser, Research Scientist at Data 61, Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia

Key References (Mostly) in Order of Appearance

PART ONE

Twenty-four years of empirical research on trust in AI: a bibliometric review of trends, overlooked issues, and future directions by Michaela Benk, et al 2024

An Integrative Model of Organizational Trust by Roger C. Mayer, James H. Davis, and F. David Schoorman 1995 The Sanction of Authority: Promoting Public Trust in AI by Bran Knowles and John T. Richards FAccT '21: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency 2021 PART TWO

Designing for Appropriate Reliance in Trust in Automation by John D. Lee and Katrina A. See Human Factors and Ergonomics Society 2004

A New Scale for the Measurement of Interpersonal Trust by Julian Rotter Journal of Personality 1967

PART THREE

A Novel Method for Trust-Sensitive Design: Applying Conjoint Analysis to Machine-Assisted Genome Annotation by Melanie McGrath, et al (including Andreas Duenser) Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems 2025 Measuring trust in artificial intelligence: validation of an established scale and its short form by Melanie McGrath, et al (including Andreas Duenser) Frontiers in Artificial Intelligence 2025 Task Load Index by Sandra Hart NASA Human Performance Research Group 1986 (also see this link)

Participants matter: Effectiveness of VR-based training on the knowledge, trust in the robot, and self-efficacy of construction workers and university students by Pooya Adami, et al including Gale Lucas Advanced Engineering Informatics 2023

Trust and Suspicion by Morton Deutsch Journal of Conflict Resolution 1958 The Original Prisoner's Dilemma Game distributed at 1950 Stanford University event (originally called "A Two*Person Dilemma") A.W. Tucker 1950 Mor