Graduation Year

2024

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

School of Accountancy

Major Professor

Uday S. Murthy, Ph.D.

Committee Member

Lisa Gaynor, Ph.D.

Committee Member

Mark Taylor, Ph.D.

Committee Member

Sandra Schneider-Wright, Ph.D.

Keywords

artificial intelligence, auditor reliance, complex estimate, decision aid, reliability, transparency

Abstract

Artificial intelligence (AI) is growing rapidly in the accounting field as audit firms invest heavily in technology to enhance audit efficiency and effectiveness. Existing research reveals two contrasting behaviors: algorithm aversion, where individuals are reluctant to rely on algorithms even if their recommendations are equivalent to humans, and algorithm appreciation, where individuals excessively rely on AI without exercising professional skepticism. This study investigates whether auditors’ algorithm aversion can be mitigated through interventions of providing reliability information and transparency information (explainability) about the AI’s processes in order to enhance auditor reliance on AI tools. The results indicate that the participants were more inclined to rely on an AI tool when provided with reliability information of the tool’s historical performance. However, transparency of the AI process did not significantly increase reliance, likely due to low statistical power. Further analyses suggest confidence in AI and perceived evidence quality produced by AI partially mediate the relation between AI transparency and reliance of the AI tool. These findings suggest that audit firms can enhance auditor reliance on AI tools primarily through providing reliability information.

Download

Share

COinS