The spread of artificial intelligence throughout almost every area in ordinary life, sooner or later will reach Public Administration, backed by promises of celerity, accuracy and other benefits when it comes do decisions in the Administrative State. This paper, built according to the critic-dialectic method, offers a conceptual set on AI modeling and the functions it might run in public organizations. It also maps initial sticking points the usage of AI can bring when adopted in public organizations without a critical perspective. The conclusion is the urgent need to revisit old canons in Administrative Law, as they might not be useful in their traditional content, to prevent power abuse in a scenery where AI decides, or at least, induces decision.