Melani Risma Gultom, Vita Fitria Sari

Pengaruh Persepsi Penggunaan Artificial Intelligence terhadap Kualitas Audit

Introduction

Pengaruh persepsi penggunaan artificial intelligence terhadap kualitas audit. Studi ini meneliti pengaruh persepsi kemudahan dan kegunaan AI terhadap kualitas audit pada auditor KAP di Indonesia. Menemukan kegunaan AI signifikan meningkatkan kualitas.

Abstract

This study aims to examine the influence of the perception of artificial intelligence ease of use and the perception of artificial intelligence usefulness on audit quality. Quantitative causality is the term for this type of research. A closed questionnaire was used to collect research data, which was then distributed via linkedIn, email, and whatsApp. The sampling method used was non-probability sampling, specifically convenience sampling, with a total of 121 valid respondents who were external auditors from Public Accounting Firms (KAP) in Indonesia. Because the data was obtained from easily accessible respondents, there is potential for sample bias that could affect the representativeness and external validity of the research findings. Data analysis was performed using multiple linear regression with SPSS version 26. The research results indicate that the perception of artificial intelligence usefulness has a positive and significant impact on audit quality, while the perception of artificial intelligence ease of use has a negative and non-significant impact on audit quality.

Review

This study tackles a highly pertinent topic concerning the influence of Artificial Intelligence (AI) perceptions on audit quality, an increasingly critical area given the rapid integration of technology within professional services. The authors aimed to examine how the perception of AI usefulness and AI ease of use individually impact audit quality among external auditors in Indonesia. The findings present a notable divergence: while the perceived usefulness of AI demonstrates a positive and significant impact on audit quality, the perceived ease of use surprisingly shows a negative and non-significant effect. This initial insight offers valuable, albeit complex, information for both practitioners and researchers navigating the evolving AI landscape in the auditing profession. While the research employs a quantitative causal approach and uses multiple linear regression for data analysis, which are appropriate for testing hypothesized relationships, significant methodological concerns arise regarding the sampling strategy. The study's reliance on non-probability convenience sampling, explicitly acknowledged by the authors, poses a substantial threat to the representativeness of the sample and the generalizability of the findings to the broader population of Indonesian external auditors. A sample size of 121, although often sufficient for regression, loses much of its statistical power and external validity when derived from an easily accessible, potentially biased group. This limitation critically hinders the ability to confidently extrapolate the observed relationships beyond the specific respondents of this study, affecting the overall strength of its conclusions. The contrasting results, particularly the negative and non-significant impact of AI ease of use, warrant deeper theoretical exploration and further empirical investigation. This counter-intuitive finding challenges common assumptions in technology acceptance models and suggests underlying nuances that the current research design might not fully capture, such as potential misinterpretations of "ease of use" in a complex auditing context or the presence of moderating factors. For future research, it is highly recommended to adopt more robust sampling techniques, such as stratified random sampling, to enhance representativeness and external validity. Additionally, integrating qualitative methods could provide richer, more granular insights into the perceptions and experiences that drive these observed outcomes. Addressing these methodological enhancements would significantly bolster the contribution of subsequent studies in this crucial domain.

Full Text

Comments

You need to be logged in to post a comment.