Opportunities and Limitations of Artificial Intelligence Technologies in Audit of Financial Statements 2030

Abstract

Relevance. The digital transformation of the economy and the exponential growth of the volumes of data generated by ERP systems create a new technological context for auditing activities. Traditional verification methods based on selective procedures and manual document processing face time, labor, and data coverage constraints. Under these conditions, the introduction of artificial intelligence (AI) technologies is considered as one of the key areas of audit modernization, which makes it possible to automate analytical processes, process one hundred percent of the transaction array, and generate predictive risk assessment models. However, the practical integration of AI into the audit of financial statements is hampered by methodological, legal and ethical barriers, as well as the lack of clear regulatory requirements for the verification of algorithms. The purpose of the study is to systematize the directions, assess the effectiveness and limitations of the use of AI in the audit of financial statements, as well as substantiate the conditions under which the integration of innovative technologies into the audit process contributes to improving the quality of audit evidence while maintaining the independence and responsibility of the auditor. The paper uses methods of theoretical analysis and synthesis of scientific literature, regulatory documents and practical situations of implementation of AI solutions by leading audit organizations. A system-functional approach has been implemented to classify the levels of analytics, as well as a risk-based method for assessing potential threats of algorithmic bias, a «black box» and the weakening of professional skepticism. The theoretical basis of the research was the works of foreign and domestic scientists in the field of auditing and its digital transformation, as well as research on statistical methods for detecting distortions (Benford’s law). The International Standards on Auditing (ISA 500, 570, 230) are used as the regulatory framework. The results of the study are to determine the levels of analytics (descriptive, diagnostic, predictive, prescriptive) and the corresponding technological stack. The key audit procedures subject to automation are systematized, it is shown that AI transforms the professional judgment of the auditor, increasing its evidence, but creating the risk of «automated trust». The groups of implementation effects and risk categories are highlighted. It has been established that the system-forming factor of efficiency is not the degree of automation, but a balanced combination of algorithmic tools and professional skepticism, ensured by continuous verification of models and the development of the regulatory framework.