The use of artificial intelligence in auditing and forensics

This dissertation examines the use of artificial intelligence for auditing and forensics. The first essay is a conceptual analysis, the second is quantitative and experimental. The first essay focuses on the ethics of AI. Accounting firms are reporting the use of Artificial Intelligence (AI) in their auditing and advisory functions, citing benefits such as time savings, faster data analysis, increased levels of accuracy, more in-depth insight into business processes, and enhanced client service. AI, an emerging technology that aims to mimic humans' cognitive skills and judgment, promises competitive advantages to the adopter. As a result, all the Big 4 firms are reporting its use and their plans to continue with this innovation in audit planning risk assessments, tests of transactions, analytics, and the preparation of audit work-papers, among other uses. As the applications and benefits of AI continue to emerge within the auditing profession, there is a gradual awakening to the fact that unintended consequences may also arise. Thus, this essay responds to the call of numerous researchers to explore the benefits of AI and investigate the ethical implications of the use of this emerging technology. By combining two futuristic ethical frameworks, this study forecasts the ethical implications of using AI in auditing, given its inherent features, nature, and intended functions. This essay provides a conceptual analysis of AI's practical ethical and social issues, using past studies and inferences based on the reported use of the technology by auditing firms. Beyond exploring these issues, this essay discusses the responsibility for the policy and governance of emerging technology. The second essay focuses on the use of machine learning in auditing. Fraud risk assessment is challenging for external auditors due to its complexity and because external auditors are usually the outsiders looking in. This essay examines the use of a framework that combines natural language processing and machine learning for detecting fraud red flags within corporate communication. The framework uses natural language processing to measure the temporal sentiments and emotions conveyed in corporate communication and the topics discussed that point to fraud red flags. The framework relies on machine learning to identify the temporal changes in the derived quantitative measures. When applied to a real corporate communication dataset for a firm with known financial statement fraud, the machine learning framework correctly flagged the implicated departments, demonstrating how auditors can use the framework for fraud risk assessments. Additionally, the essay validates the machine learning framework. To validate the machine learning framework, I used an expert panel of forensics experts with CPA certification. Given the same information, the expert panel expressed fraud risk assessments consistent with the machine learning framework. This second essay uses an ensemble of machine learning methods to analyze the temporal changes of the sentiments, emotions, and topics discussed by individuals within an organization to detect fraud cues. The key contribution of the second essay is that it examines how machine learning and textual analysis can be used to detect fraud risk cues in the organization before the issuing of financial statements (i.e., does not rely on elements of the issued financial statement and therefore can be used in continuous auditing). Since the methodology in this paper begins with unsupervised machine learning, this study demonstrates an automated approach to labeling a digital communication dataset for machine learning to detect fraud cues. The use of an unsupervised machine learning approach enables this framework to be generalizable in that there is no requirement for a context-specific pre-labeled dataset. However, there is an initial requirement for a fraud word list, as discussed in chapter 3. Based on a literature review by Sánchez-Aguayo et al. (2021), there is an identified gap in studies that use fraud detection, human behavior, machine learning and fraud theory. This second essay cuts across these four areas.Ph.D.Includes bibliographical reference