PUBLIC ADMINISTRATION IN THE CONTEXT OF AI IMPLEMENTATION: RISKS AND TRANSPARENCY CRITERIA IN INTERACTION WITH NON-GOVERNMENTAL ORGANIZATIONS

<jats:p>This article examines the integration of artificial intelligence (AI) into public administration and the associated risks, particularly in the context of interactions between public authorities (PAs) and non-governmental organizations (NGOs). It emphasizes the need to ensure openness, transparency, and the prevention of negative aspects of lobbying. The study analyzes the economic-organizational, institutional, and administrative aspects of AI implementation. It identifies and systematizes risks arising from the interaction between PAs and NGOs, such as data inaccuracies and biases, breaches of confidentiality, non-transparent and biased algorithms, as well as lobbying and a lack of oversight or accountability. The research aims to develop criteria for assessing the transparency of management decisions under the latent influence of lobbying. A set of criteria is proposed, covering information accessibility, the rationale behind decisions, stakeholder involvement, and the presence of control mechanisms. Particular attention is paid to accounting for transformational aspects (economic-organizational, institutional, and administrative) that affect the interaction between PAs and NGOs. These require a comprehensive approach to identifying risks and developing effective solutions for ensuring the responsible use of AI and enhancing public trust in government bodies.</jats:p>