Presentation Title

Quantifying Uncertainty

Presenter Information

Joel S. ElsonFollow

Advisor Information

Douglas C Derrick

Location

Dr. C.C. and Mabel L. Criss Library

Presentation Type

Poster

Start Date

3-3-2017 2:15 PM

End Date

3-3-2017 3:30 PM

Abstract

Many of us interact with automated agents every day (e.g., Microsoft's Cortana, Apple’s Siri, Amazon’s Alexa, etc.), and decision-makers at all levels of organizations utilize automated systems that are designed to enable better, faster, and more effective decisions. Understanding the conditions under which humans trust and rely upon automated agents recommendations is important, as trust is one of the mechanisms that allows for humans to interact effectively with a variety of teammates. Reliance and trust in automated systems is changing the way we process information, make decisions, and perform tasks. We conducted an experiment to determine the conditions and personality characteristics that affect human-machine interactions. Our analysis focused on the use of an automated decision aid in conditions of uncertainty. We also looked to see how perceptions of an automated decision aid’s ability related to human trust. Last, we explored how extraversion, a broad factor that encompasses the tendency to be energetic, affiliative, and dominant, related to perceptions of trust in the automated agent. We observed that in conditions of uncertainty, human decision outcomes moved in accordance with the recommendation of the agent. In addition, we found a correlation between perceptions of ability and user trust in the automated agent.

COinS
 
Mar 3rd, 2:15 PM Mar 3rd, 3:30 PM

Quantifying Uncertainty

Dr. C.C. and Mabel L. Criss Library

Many of us interact with automated agents every day (e.g., Microsoft's Cortana, Apple’s Siri, Amazon’s Alexa, etc.), and decision-makers at all levels of organizations utilize automated systems that are designed to enable better, faster, and more effective decisions. Understanding the conditions under which humans trust and rely upon automated agents recommendations is important, as trust is one of the mechanisms that allows for humans to interact effectively with a variety of teammates. Reliance and trust in automated systems is changing the way we process information, make decisions, and perform tasks. We conducted an experiment to determine the conditions and personality characteristics that affect human-machine interactions. Our analysis focused on the use of an automated decision aid in conditions of uncertainty. We also looked to see how perceptions of an automated decision aid’s ability related to human trust. Last, we explored how extraversion, a broad factor that encompasses the tendency to be energetic, affiliative, and dominant, related to perceptions of trust in the automated agent. We observed that in conditions of uncertainty, human decision outcomes moved in accordance with the recommendation of the agent. In addition, we found a correlation between perceptions of ability and user trust in the automated agent.