ARTIFICIAL INTELLIGENCE FOR FRAUD RISK ASSESSMENT IN AUDITING

Abstract

This study examines the role of Artificial Intelligence (AI), specifically ChatGPT, in enhancing auditors’ fraud risk assessments by mitigating the dilution effect. The dilution effect occurs when irrelevant information reduces the weight of relevant diagnostic evidence, thereby lowering the accuracy of auditors’ judgments. Using a 2×2 between-subjects experimental design, data were collected from 66 professional auditors enrolled in IAPI’s PPL program. The independent variables were information type (relevant vs. relevant + irrelevant) and AI use (without vs. with ChatGPT). Independent t-tests were used to compare group means, while a two-way ANOVA was used to test for interaction effects. Results show that the dilution effect significantly decreased auditors’ fraud risk assessments when irrelevant information was present without AI support. However, when ChatGPT was used, differences between relevant-only and combined conditions became insignificant. This indicates that ChatGPT effectively neutralized the dilution effect, helping auditors focus on relevant evidence. These findings highlight the potential of AI to mitigate cognitive biases in auditing and enhance decision-making accuracy. The study provides empirical contributions by integrating behavioral auditing research with AI applications, while also offering practical implications for auditors, firms, and regulators in strengthening fraud detection within the digital transformation era.

Ähnliche Arbeiten