The issue of transparence and liability for use of artificial intelligence in public administration

Cet article analyse les conséquences de l’utilisation croissante de l’intelligence artificielle (IA) dans l’administration publique, en mettant l’accent sur la transparence, la responsabilité et la protection des droits des citoyens. En s’appuyant sur le cadre juridique polonais, l’étude explore les usages actuels de l’IA dans la prise de décision administrative ainsi que les risques qui en découlent, notamment lorsque ces décisions sont opaques ou difficiles à contester. L’article souligne l’importance fondamentale de l’accès à l‘information publique, de la supervision efficace des autorités publiques, ainsi que de l’établissement de responsabilités claires dans les processus impliquant des systèmes d’IA. Il examine également les récentes évolutions réglementaires européennes, en particulier le règlement européen sur l’IA (AI Act), et met en lumière des outils tels que les évaluations d’impact algorithmique, susceptibles de renforcer la transparence. Les auteurs soutiennent qu’en l’absence de garanties solides et de législations nationales complémentaires, l’intégration de l’IA dans la gouvernance publique pourrait saper la confiance des citoyens et fragiliser les principes démocratiques fondamentaux.

This article examines the implications of the increasing use of artificial intelligence (AI) in public administration, with a particular focus on transparency, accountability, and the protection of citizens ’rights. Centered on the Polish legal framework, the study explores current applications of AI in administrative decision-making and the associated risks— especially when such decisions are opaque or difficult to contest. It underscores the critical importance of access to public information, effective oversight of public authorities, and the establishment of clear lines of accountability for AI-driven processes. The paper also evaluates recent European regulatory developments, notably the EU AI Act, and highlights mechanisms such as algorithmic impact assessments that can enhance transparency. The authors contend that, in the absence of robust safeguards and complementary national legislation, the integration of AI into public governance risks eroding public trust and undermining foundational democratic principles.

In the era of accelerating progress and modernization of everyday life, new solutions implemented increasingly eagerly at the level of state institutions and the public sector have begun to make their way into it. Changes aimed at streamlining and improving the efficiency of tasks previously relying solely on people, which included their strengths, but also their limitations, have gained momentum and have begun to be implemented ever more frequently in the operations of state administrative bodies. Forecasts from leading experts predict that in the near future, less complex tasks will be performed exclusively by artificial intelligence systems. The Polish public administration is already using artificial intelligence systems in an unsupervised manner when performing simpler and schematic tasks and this trend is noticeable in the literature. This state of affairs raises concerns from the point of view of transparency, liability issues and protection of the fundamental rights of citizens for decisions made by public administration bodies. The result of this process are tools that seek to place increasing reliance on artificial intelligence solutions in the decision-making process of public administration bodies and performance of their duties. When observing such trends, one must ask whether such solutions have a chance to effectively implement the basic principles of the rule of law, of which one of key elements is the transparency of the activities of state bodies.

In this publication, the authors analyze the essence of transparency in a democratic state under the rule of law and its potential threats from the implementation of artificial intelligence-based solutions in the public sector. In their work, the authors provide a condensed analysis of the position and importance of transparency in the organization of the state, its essence from the perspective of access to public information, supervision and control over the activities of state administrative bodies. The authors address the issues of liability of authorities for decisions made with the use of artificial intelligence and the difficulties of its enforcement in view of the complexity of such systems. The authors aim to identify problem areas in the application of artificial intelligence-based solutions from the perspective of their explainability, as well as to summarize and identify spheres burdened with gaps in both the law itself and general mechanisms for viable supervision and verification of the systems used.

Although there are many approaches to AI, in this paper, the authors have adopted the definition endorsed in the Artificial Intelligence Act by the European Parliament – which was passed by the European Parliament on 13 March 2024 and published in the Official Journal of the European Union on 12 July 2024, which came into force on 2 August 2024 (however similar to the General Data Protection Regulation, its applicability will be delayed) – as the definition of artificial intelligence when discussed in this article. This definition states that “AI system” means a machine-based system designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments’. The paper analyses the problem mainly from the perspective of the Polish legal order and it is based on the legal and actual situation of the use of artificial intelligence systems by Polish public administration bodies. However, in order to provide the most reliable and comprehensive analysis, the authors verified postulated solutions and reports from various European countries. Authors for the purpose of this paper conclude public administration to be understood as a set of activities, organizational and executive undertakings carried out for the realization of the public interest by various entities, bodies and institutions on the basis of the law and in the forms prescribed by the law, or a system composed of people, organized for the purpose of continuous realization of the common good as a public mission consisting mainly (but not exclusively) in the day-to-day execution of law, equipped for this purpose with state authority and material and technical means.