Do I want to work with this AI?
Loading...
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Alternative Title(s)
A closer look at an individual’s personality, AI advice framing, and historical AI performance in an unrelated task
Abstract
As many economies transition to Industry 4.0, Artificial Intelligences (AIs) permeate the workflows of countless organizations. While organizations often implement AIs as decision support systems, their employees frequently seem to fail at fully leveraging the potential of their new decision aids. In an attempt to help organizations reap more of the benefits of AIs as decision support systems, research has started to investigate antecedents of the willingness to collaborate with AIs. However, important questions regarding these antecedents remain and this dissertation aspires to reduce the number of open questions.
Data from 3,169 participants across five experiments offer insights into four areas. Firstly, findings align with theoretical expectations that personality, as defined by the Big Five personality traits, influences use of AI advice. Secondly, while persuasive message framing positively impacts use of algorithmic advice, customizing the AI’s message to match an individual’s personality (regulatory focus) does not alter the effect. Thirdly, individuals show a consistent preference for AI over human collaboration, regardless of the counterparts’ historical relative performance compared to the individual in an unrelated task. Additionally, superior AI performance relative to the individual in an unrelated task decreases willingness to collaborate. Fourthly, the type of support for an opponent – be it a human, a non-human-like AI, or a human-like AI – does not influence a focal person’s competitive irrationality.
My work bears considerable implications for both theory and practice. Regarding the former, my findings expand literature by experimentally testing and empirically supporting new antecedents of the willingness to collaborate with AIs and for the actual collaboration with AIs. Regarding the latter, my findings can help organizations and economies reap more of the large potential of AIs as decision support systems for humans. Lastly, I stimulate various future research avenues that may shine more light on this work’s important focus.
Description
Table of contents
Keywords
Human-AI collaboration, Personality, Performance, Competition, Behavioral experiments
Subjects based on RSWK
Mensch-Maschine-Kommunikation, Künstliche Intelligenz, Unternehmen, Mitarbeiter