Kunstig intelligens (AI) og revision: En systematisk litteraturgennemgang af teknologiens betydning for revisionsfaget

Abstract

Abstract Purpose: This study aims to investigate the impact of artificial intelligence (AI) on the audit process, the auditor and audit quality. Design: This study used a systematic literature review combined with a thematic analysis to examine the research topic. This was done through the inclusion of 108 peer-reviewed articles and an bibliometric analysis of the included literature. Furthermore, this study incorporated several theoretical frameworks to further provide analytical depth to this study, and a total of 10 different theoretical perspectives or theories were included to ascertain the impact of AI on the auditing profession. Findings: The findings were highlighted through the development of 5 specific and central themes: Audit Quality and Reliability, Efficiency and Automation, Auditor Judgment and Skepticism, Barriers to Implementation, Ethics, Regulation, and Governance. The main findings show that AI could impact the audit quality by enabling full-population testing, improved anomaly detection, fraud prevention, enhanced accuracy, improved reliability and trust, and continuous auditing. However, potential problems were identified, and include a lack of explainability when using AI, possible reduced perceived quality, stakeholder skepticism, and a possible overreliance on quantitative metrics. The possible efficiency gains include time savings, workflow automation, scalability, continuous monitoring, real-time risk alerts, increased audit coverage, and a possible reallocation of human effort to analytical and strategic tasks. However, the potential problems include, tool overtrust, deskilling of auditors, reduced client interaction, digital divide and inequality, high initial costs and uneven access across firm sizes. The impact on auditor judgment and skepticism showcased that AI could provide enhanced judgment with decision support tools, pattern recognition, improved professional insights, reinforcement of skepticism through comparative diagnostics and risk profiling, however problematic areas include automation bias, false confidence in artificial generated conclusions, weakening of professional skepticism, cognitive overload from dashboards, and uncritical acceptance of outputs. Several components related to implementation were revealed: Strategic modernization being crucial, and the potential efficiency gains were dependent on proper adoption, competitive advantage for early adopters, institutional legitimacy, however cultural resistance to change was shown, as well as a possible lack of internal capabilities for companies, and some companies merely adopted the technology as a symbolic adoption, and caused a misalignment of tools with firm strategy, and the technology showed high integration complexity. These findings show the complexity of the technology and its inclusion into the auditing profession, highlighting both opportunities, and challenges. Originality/value: This study contributes to existing knowledge of digital transformation and the impact on the auditing profession and attempts to improve the existing literature by synthesizing, analyzing, interpreting, and discussing what impacts the technology has on the auditing profession by following a structured and transparent process, and this study’s contribution can be seen through its structured analysis and overview of a fragmented research field encompassing of a wide array of theories, methodologies and focus areas.

A master thesis from Aalborg University

Kunstig intelligens (AI) og revision: En systematisk litteraturgennemgang af teknologiens betydning for revisionsfaget

[Artificial Intelligence (AI) and Auditing: A Systematic Literature Review of the Technology's Impact on the Auditing Profession]