Detailed overview
Poland is regulated by the EU AI Act and is preparing national implementation legislation. The Polish Council of Ministers accepted a draft Act on AI systems on 31 March 2026. Official materials state that the draft law is intended to ensure the application of the EU AI Act in Polish law and to create effective AI supervision in Poland. The draft is proceeding to further work in the Sejm.
The Polish draft does not replace the EU AI Act. The substantive AI classifications remain EU-based: prohibited AI, high-risk AI, transparency-risk AI, lower-risk AI and general-purpose AI models. In practice, a Polish AI compliance review must still begin by identifying whether the system is prohibited, high-risk or subject to transparency duties under the EU AI Act.
The draft law would create the Commission for the Development and Security of Artificial Intelligence, or KRiBSI, as the national authority responsible for supervision of AI rules. KRiBSI would conduct administrative proceedings, issue decisions, impose administrative sanctions, coordinate supervision of AI systems in Poland, manage regulatory sandboxes, conduct education and information activities, cooperate with EU and international bodies, and issue opinions on application of the EU AI Act.
The draft framework also includes complaint and control mechanisms. Official materials describe a model under which a person may complain to KRiBSI if they believe an AI system has operated incorrectly, unfairly or discriminatorily. The Commission would be able to inspect the system and, in urgent cases, order withdrawal or suspension of risky systems.
Penalties for substantive AI Act breaches are based on the EU AI Act. Since Poland's national law is still a draft, Polish-specific enforcement mechanics, procedures and institutional details should be assessed once the statute is finally adopted and published.
Practical requirements & details
Sourced from Regulation (EU) 2024/1689 (the AI Act) and the draft Act on AI systems accepted by the Polish Council of Ministers on 31 March 2026.
EU AI Act core duties (in Poland)
- Prohibited AI — banned.
- High-risk AI — risk management, data governance, documentation, transparency, oversight and conformity assessment.
- Transparency-risk AI — disclosure duties.
- GPAI models — documentation, transparency and copyright-policy duties.
Polish national implementation
- Draft Act on AI systems accepted by the Council of Ministers on 31 March 2026 — proceeding to the Sejm.
- KRiBSI — Commission for the Development and Security of AI as the national supervisory authority.
KRiBSI powers (draft)
- Conduct administrative proceedings, issue decisions, impose administrative sanctions.
- Coordinate supervision of AI systems in Poland.
- Manage regulatory sandboxes, conduct education and cooperation, issue opinions on EU AI Act application.
- In urgent cases — order withdrawal or suspension of risky systems.
Complaints
- Persons may complain to KRiBSI about AI systems operating incorrectly, unfairly or discriminatorily.
Penalties
- EUR 35m / 7% of worldwide annual turnover — prohibited AI.
- EUR 15m / 3% — many other AI Act operator obligations.
- EUR 7.5m / 1% — incorrect, incomplete or misleading information to authorities.
- Polish-specific enforcement mechanics will be set by the implementation Act once adopted.
Related entries
See also the European Union entry, which covers the EU AI Act (Regulation (EU) 2024/1689) — the substantive framework that this jurisdiction implements and supervises domestically.