PROSPECTS FOR USING ARTIFICIAL INTELLIGENCE AND BIG DATA IN FORENSIC ACCOUNTING OF ACCOUNTING TRANSACTIONS

Abstract

The article conducts a comprehensive scientific study of the transformational processes in the field of forensic accounting expertise, driven by the rapid diffusion of Artificial Intelligence (AI) technologies and Big Data analytics tools into financial monitoring systems. The relevance of the chosen topic is justified by the objective necessity to modernize classical methods of expert examination, which, in the context of total digitalization of accounting processes, demonstrate limited effectiveness due to their selective nature and high sensitivity to human error. The author provides a detailed retrospective analysis of the evolutionary stages of expert activity, highlighting the transition from traditional paper-based auditing (Stage 1.0) through algorithmic automation (Stage 2.0) to intelligent systems of cognitive analysis (Stage 3.0).

Special attention is paid to the methodology of implementing Machine Learning algorithms for identifying latent financial anomalies. The paper thoroughly discloses the architecture of interaction between Big Data analytics and accounting operations, which allows for the detection of complex fraud schemes, such as "mirror" transactions, fictitious counterparty relations, and double-entry bookkeeping, by analyzing unstructured data (metadata, IP logs, geolocation tracks). Based on the conducted analysis and the presented comparative tables, it is proved that the use of artificial intelligence increases the accuracy of identifying accounting risks to over 90%, which remains unattainable using traditional methodologies.

The analytical part of the article includes an assessment of the prospects for integrating cloud technologies and distributed ledger systems (Blockchain) as guarantors of the authenticity and immutability of the evidence base. Strategic vectors for further scientific research are outlined, in particular, the development of the "Explainable AI" (XAI) concept, aimed at legitimizing algorithmic conclusions in the judicial and procedural field. The conclusions emphasize that improving forensic accounting expertise through the prism of intelligent technologies transforms it from a retrospective investigative tool into a system of preventive anti-crisis monitoring. The article is of practical value to forensic experts, auditors, financial security specialists, and researchers involved in the digitalization of accounting and law.

Introduction. The current stage of global information society development is characterized by the total convergence of economic processes and high technologies. The rapid transition of business entities to the use of cloud-based accounting systems, intelligent ERP complexes, and decentralized ledgers has fundamentally transformed the financial control landscape. However, alongside the undeniable benefits of digitalization, new and technologically sophisticated types of economic offenses are emerging. Modern financial fraud is increasingly masked as legitimate transactions created through algorithmic manipulations, which are virtually impossible to identify using the classical toolkit of forensic accounting expertise.

Traditional expert research methodology, based on the visual verification of primary documents and retrospective analysis of limited samples, demonstrates signs of conceptual exhaustion in today's realities. In the context of Big Data generation, where the daily volume of operations can reach millions, the expert economist faces the challenge of "information overload." This creates an urgent need for the integration of Artificial Intelligence (AI) as a cognitive assistant capable of continuous monitoring and instantaneous identification of anomalies, which determines the high relevance of this scientific research.

Purpose. This paper aims to explore the prospects of integrating AI and Big Data technologies into forensic accounting procedures for accounting operations.

Materials and methods. The research materials include academic works on forensic accounting, digital forensics, Big Data, and AI, as well as regulatory provisions governing evidence and forensic activities. The methodological basis of this scientific work is a systematic approach to the study of transformational processes in forensic activities. To ensure the objectivity and reliability of the obtained results, a complex of general scientific and special methods was used, which allowed for the analysis of the problem from various perspectives — from technical-technological to procedural-legal. To solve the set tasks, the following methods were applied:

• the method of historical and logical retrospection was used to build an evolutionary model of forensic expertise development (from stage 1.0 to 3.0), which allowed for the identification of patterns in the transition from analog to cognitive methods of information processing;
• comparative analysis served as the basis for contrasting traditional manual control techniques with automated Big Data systems. This specific method made it possible to clearly demonstrate the gap in anomaly detection efficiency indicators (as reflected in Tables 1 and 2);
• the method of deductive modeling was applied to describe the mechanisms of neural network operation in the process of identifying latent connections between fictitious counterparties and primary accounting registers;
• statistical methods and the generalization method were used in the analysis of quantitative indicators of AI prediction accuracy and in forming conclusions regarding the probability of detecting various categories of offenses (specifically, double-entry bookkeeping and misappropriation of funds).

The applied combination of methods allowed not only for the confirmation of the advantages of digital technologies but also for a critical assessment of the barriers to their implementation, ensuring a comprehensive and unbiased study of the chosen subject matter.

Results. In the course of the study, it has been substantiated that the integration of Artificial Intelligence and Big Data into forensic accounting processes enables a transition from selective testing to continuous intelligent monitoring of 100% of accounting transactions. It has been established that the use of machine learning algorithms provides an increase in the efficiency of detecting hidden financial anomalies, particularly fictitious counterparty relations and double-entry bookkeeping, to a level of 85–92%, which critically exceeds the capabilities of traditional methodologies.

 It has been proven that the digital transformation of expert activities leads to a shift from reactive investigation to real-time predictive risk modeling. The evolutionary model and the architecture of intelligent research developed by the author confirm that the object of modern expertise is no longer limited to a primary document but encompasses a set of digital footprints and metadata. This minimizes the influence of the human factor and ensures maximum objectivity of the evidence base in the judicial process.

Discussion. A promising direction for further research is the development of unified industry standards for the use of "Explainable AI" (XAI), which will allow for the transformation of complex mathematical Big Data correlations into a transparent and legally understandable evidence base for judicial proceedings. Further convergence of blockchain-based smart contracts with predictive machine learning algorithms opens up opportunities for creating systems of continuous intelligent auditing. This approach will allow for shifting the vector of expert activity from stating the facts of past offenses to the strategic prevention of economic crimes by automatically blocking anomalous transactions in real time.