TRUST IN MACHINE AGENTS UNDER REALISTIC THREAT

Abstract

Measures of actual human-machine interaction are required to generate ecologically valid, translatable discoveries that enhance human-systems integration and performance. Here, the proposed methods center on dyadic human-robot decision tasks to assess key determinants of human trust in robot recommendations (e.g., factors such as threat-salience and/or robot appearance), and the concomitant impacts of trust on performance. We will assess the impact of psychological representations of the competence, benevolence and integrity of robotic partners as these representations predict behavioral trust outcomes with regard to navigation, threat-detection, and the use of force. In addition to these laboratory studies, parallel online studies will be conducted to assess perceptions of, and conformity with, screen-mediated software agents.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 12, 2021
Source ID
FA95502010347

Entities

People

  • Colin Holbrook

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of California

Tags

Readers

  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • Autonomy
  • Autonomy - Human-Robot Interaction