Transforming Auditing with Artificial Intelligence: A Framework for Fraud Detection and Responsible Adoption

Abstract

— This study investigates how Artificial Intelligence (AI) can transform auditing by improving fraud detection and audit quality. However, traditional methods, limited by sampling and manual inspection, failed to detect complex frauds such as Luckin Coffee. Using the Technology – Organization – Environment (TOE) framework and Socio-Technical Systems (STS) theory, the research adopted a qualitative case study design. Audit failures were reconstructed through triangulated evidence and counterfactual reasoning, assessing how Machine Learning (ML) and Natural Language Processing (NLP) could have detected anomalies in real time. Findings show AI provides broader coverage, timeliness, and granularity while preserving professional skepticism. The study contributes by situating AI adoption within socio-technical and organizational contexts and proposing a framework for responsible implementation, and it is crucial to position AI as a tool that complements, rather than substitutes, human auditors.