USE OF ARTIFICIAL INTELLIGENCE IN ELECTRONIC PUBLIC PROCUREMENT SYSTEMS: LEGAL ASPECTS, CHALLENGES AND PROSPECTS

Abstract

The article provides a thorough legal analysis of the integration of artificial intelligence (hereinafter - AI) technologies into the public procurement sector in Ukraine. Given the rapid development of digital tools in the field of public administration and the functioning of the ProZorro electronic system, the author emphasizes the relevance of regulatory support for the use of algorithms in procurement procedures. Particular attention is paid to the issues of legal certainty, algorithmic transparency, ensuring the right to appeal against automated decisions, and personal data protection.

Against the backdrop of martial law, the digitalization of public procurement is not only of managerial but also of security importance: the effective functioning of the ProZorro electronic system is a guarantee of transparent use of budget funds and maintaining trust in the state and international donors. At the same time, the use of algorithmic mechanisms without proper legal regulation creates a number of legal threats: from biased or discriminatory decisions to restrictions on participants' access to justice in case of erroneous automatic evaluation.

The author analyzes the current legislation of Ukraine in the field of public procurement, including the Law of Ukraine «On Public Procurement» and resolutions of the Cabinet of Ministers of Ukraine. It is established that the legislative framework does not contain sufficient guarantees of transparency of automated decision-making, human-in-the-loop mechanisms, or procedures for appealing such decisions. There is also a lack of provisions on liability for the results of AI analysis in public procurement.

At the same time, the article provides a comparative analysis of international practice. In particular, the experience of France (Chorus Pro), Canada (Responsible AI Framework), the United Kingdom (A Guide to Using AI in the Public Sector), Estonia (integration with registries), Germany (right to human review of decisions), and the EU in the context of the AI Act is considered. These examples allowed us to formulate guidelines for harmonizing Ukrainian legislation with European and global approaches.

In addition, the subject of the study of the outlined issues was the possible risks of introducing AI into the public procurement system. In particular, the author highlights the following: non- transparency of algorithms, discriminatory effects due to bias in training data, lack of accountability mechanisms and appeal procedures, and the possibility of abuse in wartime due to formal “algorithmic” exclusion of unwanted participants. Special attention is paid to personal data protection and requirements for explainable AI.

The final part proposes a number of legislative initiatives: supplementing the Law of Ukraine “On Public Procurement” with provisions on AI, introducing a code of ethics for the use of AI in the public sector, creating an independent body to control algorithms, mandatory algorithmic legal assessment (LAIA), and changing the procedure for appealing decisions.

The proposed changes are aimed at ensuring a balance between innovative efficiency and compliance with legal and ethical standards. The article is of interest to lawyers, civil servants, IT professionals, and researchers of digital governance.