Concepedia

TLDR

As society increasingly relies on automated systems, understanding operator trust—often misrepresented by anthropomorphic scales—is essential for effective use. The study aims to validate the Trust of Automated Systems Test (TOAST), a scale that avoids anthropomorphizing and measures trust dimensions. The authors conducted two validation studies of the nine‑item TOAST scale, assessing its factor structure and reliability. Confirmatory factor analyses confirm that the nine‑item TOAST reliably measures operator trust in both military and civilian automated systems.

Abstract

This lecture discusses the validation of a new scale, the Trust of Automated Systems Test (TOAST). As we increasingly rely on automated systems to perform tasks, it becomes ever more important to understand the ways in which real operators will use these systems. Trust is a critical determinant of this use. Previous efforts to translate trust, typically an interpersonal human experience, into the domain of artificial cognition and inanimate objects, has met with mixed success. Existing scales tend to anthropomorphize systems and attribute intent to relatively simplistic decision processes. This anthropomorphization has led to operator dissatisfaction with these scales and, consequently, poor data quality. In two studies, I present the validation effort for a scale that avoids anthropomorphizing while still assessing the theoretically-driven underlying dimensions of trust formation. Confirmatory Factor Analyses provide evidence that the TOAST, a nine-item scale measuring operator Understanding of the system and belief in its Efficacy is a valid metric of operator trust of both military and civilian systems.