Me, Myself, and AI: Investigating the Influence of Self-Esteem on Adoption of Human vs. AI Suggestions

Utrecht University 2024

With the increasing use of Artificial Intelligence (AI) in many domains, collaborations between humans and algorithms are becoming more common. Several factors have been identified that influence an individual’s receptivity to the suggestions of AI agents. However, the impact of many of these variables can differ across contexts. In human-human relationships, evidence suggests that global self-esteem predicts an individual’s likelihood to accept the advice of others. As self-esteem is considered to be stable across situations, this study sought to replicate this effect with both human and artificial teammates, and investigate if the relationship between these variables is stronger when the suggestions come from a human than an AI algorithm. The bin packing game was utilized as the experimental task and participants reported their likelihood to adopt alternative solutions to the game. Contrary to hypotheses, regression analyses revealed no significant relationship between self-esteem and suggestion acceptance. Exploratory analyses suggest that these variables may have demonstrated a positive relationship in the AI group analyzed alone, but not in the human group. These results are not in line with previous findings regarding advice-taking in human-human relationships and bring into question the stability of this relationship.

This project investigated the potential influence of individual self-esteem level on likelihood to accept advice from AI and human teammates. As self-esteem is considered fairly stable across contexts and has previously been shown to predict advice acceptance in human-human relationships, we sought to test if this effect extends to human-AI relationships. Participant experiments were conducted using the bin-packing task. Surprisingly, the primary results were not in line with the hypotheses.; Master Thesis; Student Thesis

Open access content. Open access content
CC-BY-NC-ND
Open Access

EN

QGJ oai:studenttheses.uu.nl:20.500.12932/46788
https://studenttheses.uu.nl/handle/20.500.12932/46788
1456106546

UNIVERSITEIT UTRECHT
From OAIster®, provided by the OCLC Cooperative.

edsoai.on1456106546