Date of Award

12-24-2025

Date Published

1-16-2026

Degree Type

Dissertation

Degree Name

Doctor of Professional Studies

Department

Information Management

Advisor(s)

Jaime Banks

Keywords

Anthropomorphic Design;Anthropomorphism;Generative AI/LLMs;Knowledge Work;Self-similarity;Trust

Abstract

Research on perceived trustworthiness towards large language models (LLMs) has paid limited attention to the psychological mechanisms that give rise to trust itself, leading to varying conclusions on how interface design elements affect user perceptions of trustworthiness. The initial intent of this study was to test a moderated mediation model in which perceived anthropomorphism mediates the development of perceived trustworthiness towards LLMs, conditional on perceptions of self-similarity, among knowledge workers. However, results from a pilot study did not empirically support the inclusion of self-similarity as a moderator. As a result, this research was refined to a simple mediation model, where I examined how perceived anthropomorphism functions as a key psychological mechanism that links the anthropomorphic design of an LLM to perceptions of trustworthiness in LLM-supported knowledge work. An experiment was conducted (n=161) to test two simple mediation models, each addressing how perceived anthropomorphism affects the performance and moral dimensions of perceived trustworthiness by an LLM with varying levels of anthropomorphic design. Results from both models revealed full mediation, indicating that the anthropomorphic design of the LLM affected participants’ perceived trustworthiness, entirely through participants’ subjective sense of the LLM’s humanlike qualities (i.e., perceived anthropomorphism). Findings contribute to existing literature by empirically demonstrating that perceptions of anthropomorphism serve as a core psychological mechanism through which perceived trustworthiness in LLM emerges. Implications and opportunities for future research are discussed.

Access

Open Access

Available for download on Sunday, January 16, 2028

Share

COinS