Abstract

The progressive evolution of the use of It in the exercise of administrative power has invested the public administration and the public decision-making process, redesigning its form and substance, influencing organisation and procedures. The automation of administrative proceedings not only allows to redesign internal processes to increase efficiency and improve organizational performance, but also changes the very nature of the proceeding: cancelling the traditional sequence of acts, flattening all the phases, rendered contextual, thus determining the end of the administrative proceeding as known up to now. The use of Artificial intelligence (Ai), which is often seen as a panacea to the problems of the Pa, must not confuse the neutrality of the instrument (enough to credit it) with the neutrality of the method. Indeed, the Ai has itself become the subject of decisions, in administrative proceedings, as in the legal system itself. It is therefore crucial to analyse the design of algorithms, with regard to the opacity of design and transparency on the values they incorporate, in particular with regard to possible unequal treatment resulting from Machine learning (Ml). The analysis of concrete cases highlights the possible repositioning of the different actors involved in the automated administrative proceeding, compared to the redesign of public services and participation in decision-making processes.