Poland

[PL] Act on artificial intelligence systems

IRIS 2026-8:1/25

Agnieszka Grzesiok-Horosz

University of Silesia in Katowice

The Act of 3 July 2026 on artificial intelligence systems was published in the Journal of Laws on 27 July 2026. The full launch of the national system of supervision over artificial intelligence (AI) is proceeding in stages. A number of its provisions had already come into force on 28 July, however, a fundamental part of the act became applicable on 11 August 2026 and another part will come into force on 28 October 2026.

The act is designed to establish a legal framework for comprehensive support of innovations in the area of AI, and to transpose the EU AI Act. The Polish act ensures the application in Poland of solutions adopted under the EU regulation establishing uniform rules on AI systems throughout the European Union, ensuring the protection of fundamental rights and citizens’ safety when using AI systems, increasing public confidence in new technologies, and creating favourable conditions for the development of AI technologies.

Supervision over the application of the AI Act and the Act on artificial intelligence systems will be exercised by a dedicated body – the Commission for the Development and Safety of Artificial Intelligence (Komisja Rozwoju i Bezpieczeństwa Sztucznej Inteligencji – KRiBSI), which is going to obtain administrative support from the Ministry of Digital Affairs. As an independent authority, the Commission will be presided over by its chair, appointed by the Sejm (the lower house of the Polish Parliament) with the Senate’s consent, for a five-year term. The first appointment will be made in October 2026, and the Commission’s activities can be initiated in November 2026. Under the above-mentioned act, the Commission will be a collegiate body, composed of representatives of: the President of the Office of Competition and Consumer Protection, the Polish Financial Supervision Authority, the National Broadcasting Council, and the President of the Office of Electronic Communications. This solution is supposed to ensure an effective response to the challenges involved in the development of AI while drawing, at the same time, from the best experiences of the existing regulatory bodies. The Social Council for Artificial Intelligence will form the consultative and advisory body, composed of 9 to 15 members appointed for a two-year term.

The competences of the new regulator will be extensive. KRiBSI will examine any reported irregularities in the operation of AI systems (e.g. causing health or safety risks, making a wrong decision) and, if necessary, react to such flaws. As of 28 October 2026, citizens, businesses and institutions will be able to submit complaints to the Commission regarding the operation of AI systems. A natural person, legal person or organisational unit will be able to point to an AI system, provide essential facts, and explain why they think the provisions of the AI Act or of the Polish act have been violated.

KRiBSI is supposed to support the development of an AI ecosystem by creating and managing regulatory sandboxes, conducting educational and informational activities, supporting innovations and scientific research.

The Act on artificial intelligence systems introduces a legislative framework for the operation of regulatory sandboxes in Poland as one of the first countries in Europe to do so. Such regulatory sandboxes allow businesses and institutions to test new solutions in a safe and controlled environment, under the supervision of a competent authority and according to specific regulatory requirements. These sandboxes are intended to speed up the development and commercial utilisation of the technologies discussed while ensuring, at the same time, supervision over the key issues of safety and ethics. Polish businesses, startups and research institutions will develop modern technologies in compliance with legal requirements. It should be emphasised that micro, small and medium-sized enterprises will be in a position to participate in the regulatory sandboxes free of charge.

KRiBSI will also issue individual opinions on both existing solutions and planned AI implementations, helping businesses and institutions feel confident that their solutions comply with regulations before adopting them within their structures. In principle, the Commission should answer an enquiry within 30 days, and in particularly complicated matters, within 60 days. Businesses and institutions will provide their own legal assessment and ask the regulator if the assessment is correct. A lack of response within the prescribed deadline shall constitute approval of the enquiring party’s position by the regulator. After anonymisation, individual opinions and explanations are to be published and made available to the entire market. This is seen as an opportunity to create a “map for practical application of the AI Act”.

Moreover, the rationale behind the new provisions is to create favourable conditions for the development of modern technologies. Under the newly introduced legislation, the Polish Minister of Digital Affairs will be able to provide financial support for scientific research and experimental development in AI.

The rules described will increase the safety of users of AIsystems and ensure favourable conditions for the implementation of modern technologies in Polish businesses and institutions.