[4H4-OS-4b] OS-4

2:45 PM - 3:00 PM JST(5:45 AM - 6:00 AM UTC)

[4H4-OS-4b-06] A Structural Equation Modeling Analysis of Mediating Factors of External Blame in Human AI Collaborative Decision Making Tasks in Human Resource ContextsThe Role of Accountability Design and Information Presentation in Third Party Evaluations

〇Keito Miyake^1,2, Seiji Yamada³ (1. The Graduate University for Advanced Studies, SOKENDAI, 2. The National Institute of Informatics, 3. Kanagawa University)

Keywords:

AI,AI Blaming,Structural Equation Model

This study examines how external blame is formed by third party observers in human AI collaborative decision making tasks in human resource contexts. Although AI is increasingly used as decision support in personnel decisions such as hiring, responsibility attribution and blame targets remain insufficiently examined. An experimental task simulating a hiring scenario was conducted, in which observers were presented with a collaborative decision making process involving a human decision maker and an AI. The human decision maker and the AI were defined as entities that provided recommendations to hire or reject candidates based on candidate evaluations. The experiment manipulated decision maker confidence display, the locus of accountability, and the presence of explainable AI. After observing the decision making process, participants evaluated external blame and related psychological factors. Structural equation modeling was employed to examine relationships among factors influencing external blame and their mediating structures. This study provides a framework for investigating how accountability and information presentation in human AI collaborative decision making affect third party evaluations in human resource contexts