The authors comparatively analyze the impact of artificial intelligence (AI) on security in public administration. The rapid advancement in AI technology offers new opportunities, but it also entails significant risks for the public sector. The analysis covers the current use of artificial intelligence in administration, the risks associated with the introduction of artificial intelligence into public administration, plans for the future and challenges in Poland, the United States and China. To ensure the safe use of AI, it is necessary to conduct a broad discussion on the scope of implementation of AI in the functioning of public administration and, on this basis, to define a regulatory framework. The topic requires an interdisciplinary approach, combining computer, legal and public policy knowledge. The article addresses the critical issue of the impact of AI systems on the security of public administration, a key problem for the responsible and effective implementation of technological innovations in the public sector. The proposed interdisciplinary approach combines artificial intelligence, legal and ethical knowledge and public policy. It can contribute to the development of comprehensive solutions for the safe use of AI in public administration, in line with the public interest.