DALL’AMMINISTRAZIONE ANALOGICA ALL’AMMINISTRAZIONE PER ALGORITMI. VARIANTI E INVARIANTI DI UN PROCESSO EVOLUTIVO

Excessive caution and distrust in the use of AI systems for decision-making may lead to the renunciation of an extremely useful technology, especially for public administrations. When addressing the issue of the use of AI by public bodies, it is better think in terms of procedure rather than algorithmic administrative provision (in the sense of a fully/partially automated decision), and to consider the technological output not as a decision with a direct impact on the citizens, but as a decision proposal to be submitted to the prior hearing of the interested parties, from which any inconsistencies, discriminations, or gaps in the functioning of the system may emerge. In the case of discretionary acts, in particular, the “prediction” made by the machine on the basis of statistical/probabilistic correlations should be considered as a suggestion to the officer, who will have to take it into consideration alongside other relevant elements and, in any case, taking into account the observations made by the interested parties. In general, the use of artificial intelligence systems within public administrations must be inspired by the criterion of complementarity, and be aimed at preserving and enhancing human capabilities, not replacing them