何が起きたか
On 28 October 2026, Articles 8-18 and Chapters 3-5, 8 and 9 of Poland's Law on AI Systems (Ustawa o systemach sztucznej inteligencji, law of 3 July 2026, published as Dziennik Ustaw poz. 1003) entered into force. These provisions give the Commission for the Development and Security of Artificial Intelligence (KRiBSI) — Poland's AI Act market-surveillance authority and single point of contact — its operational enforcement toolkit: conduct of inspections (Art. 49), post-inspection procedures and recommendations, warning notices (Art. 61), power to order cessation of an AI system with immediate effect (Art. 63), and the penalties framework. Penalties are capped by AI Act Art. 99: up to EUR 35m or 7% of worldwide turnover for prohibited practices, EUR 15m or 3% for provider/deployer obligations, and EUR 7.5m or 1% for misleading information. KRiBSI had existed on paper since August 2026 but could not previously enforce against businesses.
なぜ重要か
Poland is now a fully operational enforcement venue for the EU AI Act. Any organisation deploying AI systems in Poland (including ChatGPT/Copilot-style tools as 'deployers', HR/recruitment AI, emotion-recognition or deepfake-generating systems) is now exposed to KRiBSI inspections, administrative proceedings, and the AI Act's turnover-based fines. It materially raises the compliance stakes for AI deployers and providers operating in Poland and removes the 'paper regulator' gap the EU AI Act leaves until Member States stand up enforcement.
必要な対応
AI providers and deployers in Poland should immediately verify AI Act Art. 5 prohibited-practice compliance, Art. 50 transparency/deepfake-labelling compliance, and readiness of technical documentation and risk-management files, in anticipation of KRiBSI market-surveillance inspections (which begin via a 7-day notice under Art. 49).