DIP: 18.02.053/20251003 
DOI: 10.25215/2455/1003053 
Abstract
The integration of Artificial Intelligence (AI) into forensic accounting is transforming the landscape of fraud detection and financial investigation. Traditional forensic accounting methods often rely on manual scrutiny, which can be time-consuming, error-prone, and insufficient to detect complex or subtle financial irregularities. This paper explores how AI, particularly machine learning and anomaly detection algorithms, enhances the ability to identify fraudulent patterns hidden within vast and complex financial datasets. By focusing on data anomalies—irregularities or deviations from expected patterns—AI-driven forensic tools can uncover indicators of fraud that may elude conventional analysis.
The study examines various AI models utilized in forensic accounting, including unsupervised learning techniques like clustering and autoencoders, as well as supervised models trained on labeled fraud datasets. It evaluates the efficacy of these tools in real-world scenarios, such as corporate financial statement manipulation, insider trading, and cyber-financial crimes. Furthermore, the paper discusses the challenges of data quality, algorithmic transparency, and the potential for AI bias, offering strategies to mitigate these risks.
The research underscores the importance of human expertise in interpreting AI-generated insights, advocating for a hybrid approach where accountants and data scientists collaborate. Ethical considerations, legal implications, and regulatory frameworks are also analyzed to understand the broader impact of AI on forensic accounting practices.
AI-powered anomaly detection represents a significant advancement in forensic accounting, enabling more proactive and accurate identification of financial fraud. However, successful implementation depends on the integration of advanced technology with professional judgment, ethical standards, and continuous learning. This paper contributes to the evolving discourse on the role of AI in safeguarding financial integrity in an increasingly digitized world.
The author(s) appreciates all those who participated in the study and helped to facilitate the research process. 
The author(s) declared no conflict of interest. 
ISSN 2455-670X