LEGAL REGULATION OF THE USE OF ARTIFICIAL INTELLIGENCE IN ORGANIZATIONAL MANAGEMENT SYSTEMS: MANAGERIAL RISKS AND COMPLIANCE

Abstract

The article presents a comprehensive study of the legal regulation of artificial intelligence (AI) within corporate management systems in the context of deepening European integration. The relevance of the topic is driven by the rapid proliferation of AI technologies in business practice and the significant lag of the regulatory framework behind the pace of digital transformation at the enterprise level. The study is grounded in a systemic analysis of Regulation (EU) 2024/1689 (EU AI Act) as the foundational international regulatory instrument in this domain, complemented by a comparative legal analysis of its provisions against current Ukrainian legislation and the strategic initiatives of the Ministry of Digital Transformation of Ukraine. The article provides a detailed examination of the four-tier risk-based classification of AI systems established by the EU AI Act and analyses its practical implications for various functional divisions of an organisation — from human resources management and marketing to financial control and security systems. Three key groups of managerial risks arising from uncontrolled AI deployment are systematised: legal risks (intellectual property infringement and personal data leakage); operational and strategic risks (hallucination effects and algorithmic bias); and ethical and reputational risks (the black-box effect and erosion of employee trust). A step-by-step framework for building a corporate AI compliance system is developed and theoretically substantiated, encompassing AI tool auditing, risk assessment, internal policy regulation, technical data protection controls, and personnel training. An organisational model for distributing AI governance responsibilities is proposed, introducing the roles of AI Compliance Officer and an AI Ethics Committee composed of cross-functional senior management. A compliance risk matrix is presented as a practical decision-support tool for key business processes including HR, marketing, finance, and security. The article separately examines the specifics of Ukrainian legislative adaptation to EU AI Act requirements and identifies the principal barriers to AI compliance implementation in domestic business — namely the shortage of professionals combining IT and legal expertise, financial constraints under wartime conditions, and cultural resistance manifested as Shadow AI practices. The Estonian regulatory model is considered as a benchmark for state-supported compliance infrastructure. The scientific novelty of the study lies in the formation of an applied corporate AI compliance model adapted to the conditions of European integration and the operational realities of Ukrainian enterprises, as well as in the theoretical conceptualisation of the "human oversight paradox" in the context of automation bias among managers. The article argues for the adoption of a human augmentation model, wherein AI serves as an analytical assistant while final decision-making authority and legal accountability remain with the human manager.