Structural Dimension of Legal Regulation of AI

Abstract

this article systematises contemporary approaches to the legal regulation of Artificial Intelligence (AI). It focuses on the Ukrainian context in light of global challenges. The authors aim to outline a multi-level structure for such regulation, encompassing meta, strategic, and tactical levels, and to demonstrate the necessity of aligning AI with societal values.

The study employs an interdisciplinary approach, combining legal analysis, comparative legal methods, ethical reflection, and philosophical conceptualisation. It analyses Ukrainian legislation, international normative acts, scholarly positions, and examples of AI utilisation in public administration.

It is argued that effective legal regulation of AI should be implemented at three levels: strategic (addressing long-term challenges and harmonising with international law), tactical (regulating already implemented systems and minimising risks), and meta-level (comprehending human goals and defining the social meaning of technologies). The article addresses the issue of uncertainty surrounding human values within the context of the alignment problem and examines its significance for AI legal regulation and policy. The development of socially-oriented AI use is reviewed through the example of creating the CrimeDataLab platform for integrating open data on crime prevention. It is demonstrated how the fragmentation, methodological incompatibility, and bias of statistical data in criminal justice present significant obstacles to the advancement of AI technologies, particularly in forming reliable assessment and forecasting models.

The proposed three-level model enables the formulation of a more comprehensive AI regulation policy, ensuring responsiveness to contemporary challenges and a strategic orientation towards the public good. Ukraine has an opportunity to integrate into the global regulatory landscape by combining technological development with a value-based understanding of its objectives.