Модель аналізу безпеки фінансового сектору під впливом штучного інтелекту
Вантажиться...
Дата
2025
Автори
Назва журналу
Номер ISSN
Назва тому
Видавець
Анотація
У статті представлено організаційно-інформаційну модель аналізу безпеки фінансового сектору з урахуванням зростаючого впливу штучного інтелекту (ШІ). Визначено ключові переваги та виклики інтеграції традиційної аналітики, машинного та генеративного навчання в кредитуванні, страхуванні, управлінні активами й платіжних системах. Особливу увагу приділено ризикам: «стадній поведінці», кіберзагрозам, втраті конфіденційності, ефектам мережевої взаємозалежності, алгоритмічним змовам і модельним галюцинаціям. Водночас розкрито потенціал ШІ у виявленні шахрайства, підвищенні ефективності та прийнятті стратегічних рішень. Крім того, підкреслено необхідність прозорого регулювання, стійкості та підзвітності ШІ-моделей задля підтримки фінансової стабільності в умовах цифрової трансформації.
This article presents an organizational and information model (OIM) designed to analyze the security of the financial sector amid the accelerating integration of artificial intelligence (AI) technologies. The model encompasses three key analytical blocks: traditional analytics, machine and deep learning, and generative AI (GenAI). It captures both the benefits and systemic risks stemming from AI's implementation in core areas of financial services, including intermediation, insurance, asset management, and payment systems. The research identifies enhanced efficiency, fraud detection, risk management, and personalized services as primary advantages of AI adoption. However, it also outlines critical vulnerabilities such as model opacity ("black box" effects), data bias, privacy threats, algorithmic collusion, systemic herding behavior, and hallucinations in AI-generated outputs. The study details how various AI technologies can reduce underwriting costs, improve credit scoring using unstructured data, and optimize high-frequency trading and liquidity management. At the same time, it highlights that reliance on uniform datasets and third-party providers introduces structural homogeneity and single points of failure, increasing systemic fragility. GenAI, with its speed, autonomy, and ubiquity, further amplifies these effects by enabling automation at an unprecedented scale. The article systematically classifies AI’s potential benefits, from regulatory compliance and predictive analytics to innovation in financial product development and transformation in fintech education. Moreover, the article addresses both micro-risks – such as cyber threats, data fragmentation, and discriminatory algorithmic outputs – and macro-risks, including labor market shocks, financial market procyclicality, and cascading failures. Through analysis of past financial disruptions and current empirical data, the article underscores the need for robust, transparent, and adaptive regulatory frameworks tailored to national contexts. The proposed OIM integrates actors, data sources, evaluation criteria, analytical tools, and principles into a cohesive model, with a specific submodel focusing on the interplay between AI technologies and financial stability.
This article presents an organizational and information model (OIM) designed to analyze the security of the financial sector amid the accelerating integration of artificial intelligence (AI) technologies. The model encompasses three key analytical blocks: traditional analytics, machine and deep learning, and generative AI (GenAI). It captures both the benefits and systemic risks stemming from AI's implementation in core areas of financial services, including intermediation, insurance, asset management, and payment systems. The research identifies enhanced efficiency, fraud detection, risk management, and personalized services as primary advantages of AI adoption. However, it also outlines critical vulnerabilities such as model opacity ("black box" effects), data bias, privacy threats, algorithmic collusion, systemic herding behavior, and hallucinations in AI-generated outputs. The study details how various AI technologies can reduce underwriting costs, improve credit scoring using unstructured data, and optimize high-frequency trading and liquidity management. At the same time, it highlights that reliance on uniform datasets and third-party providers introduces structural homogeneity and single points of failure, increasing systemic fragility. GenAI, with its speed, autonomy, and ubiquity, further amplifies these effects by enabling automation at an unprecedented scale. The article systematically classifies AI’s potential benefits, from regulatory compliance and predictive analytics to innovation in financial product development and transformation in fintech education. Moreover, the article addresses both micro-risks – such as cyber threats, data fragmentation, and discriminatory algorithmic outputs – and macro-risks, including labor market shocks, financial market procyclicality, and cascading failures. Through analysis of past financial disruptions and current empirical data, the article underscores the need for robust, transparent, and adaptive regulatory frameworks tailored to national contexts. The proposed OIM integrates actors, data sources, evaluation criteria, analytical tools, and principles into a cohesive model, with a specific submodel focusing on the interplay between AI technologies and financial stability.
Опис
Ключові слова
Штучний інтелект, фінансова безпека, організаційно-інформаційна модель, генеративний ШІ, машинне навчання, фінансова стабільність, ризики автоматизації, intelligence, financial security, organizational and information model, generative AI, machine learning, financial stability, automation risks
Бібліографічний опис
Недобой С. Модель аналізу безпеки фінансового сектору під впливом штучного інтелекту [Текст] / С. Недобой // Цифрова економіка та економічна безпека : науково-практичний журнал / Причорноморський науково-дослідний інститут економіки та інновацій, Сумський державний педагогічний університет імені А. С. Макаренка ; [гол. ред. О. Ю. Кудріна, редкол.: В. В. Божкова, В. І. Борщ, Н. М. Вдовенко та ін.]. – 2025. – № 2 (17). – С. 136–141. – DOI: https://doi.org/10.32782/dees.17-22