We are pleased to invite you to read the article below and also published in Sole24Ore, by our collaborator Prof. Avv. Guido Befani.
Thanks to its advanced skills, it allows us to broaden and deepen the aspects related to the legitimacy and transparency requirements that an algorithm must have in order to comply with the legal canons of the administrative procedure.

Recourse to the algorithm must be correctly framed in terms of organizational module, procedural and preliminary instrument, subject to the typical checks of every administrative procedure, which remains the modus operandi of the authoritative choice, to be carried out on the basis of the legislation conferring power and purposes attributed by the same to the public body, holder of power. Specifically, two prominent aspects are of fundamental importance to qualify their legitimacy, as elements of minimum guarantee for any hypothesis of use of algorithms in public decision-making:
a) full knowledge upstream of the module used and of the criteria applied;
b) the imputability of the decision to the body holding the power, which must be able to carry out the necessary verification of the logic and legitimacy of the choice and the results entrusted to the algorithm.
In this context, even in the face of a knowable and understandable algorithm, which is not the only reason for the decision, it must also not take on a discriminatory nature. This is what the VI Section of the Council of State affirms, with the sentence of 4 February 2020, n. 881.

The deepening
The Council of State intervened on the legitimacy and transparency requirements that an algorithm must have in order to comply with the canons of legality of the administrative procedure.

The decision
With the decision that rekected the appeal of the Miur against the sentence of the Lazio TAR no. 6607 of 2019 (with which the illegitimacy of the procedure for assigning offices through the algorithm was declared), the Board was able to note how, in general theory, also the Public Administration must be able to exploit the relevant potential of the so-called digital revolution. For the College, in fact, the use of computer algorithms for making decisions concerning the public and private spheres would be based on the feared gains in terms of efficiency and neutrality, where already in many fields the algorithms promise to become the tool through which corrects the distortions and imperfections that typically characterize cognitive processes and the choices made by human beings, and where, in this context, the decisions taken by the algorithm thus take on an aura of neutrality, the result of aseptic rational calculations based on data. For the College, however, with respect to the evolution of the Digital Administration Code on the so-called IT administrative act, through the use of the algorithm it is not just a question of experimenting with different forms of externalization of the Administration's will, or of identifying new methods of communication between Administration and private individuals, (as in the case of citizens' participation in administrative decisions through social networks or digital platforms), because by resorting to the algorithm we are faced with a situation that, in the doctrinal context, has been effectively qualified with the expression of "revolution 4.0" which, referring to the public administration and its activities, describes the possibility that the procedure for forming the administrative decision is entrusted to a software, in which a series of data is entered so as to arrive, through the automation of the procedure, to the final decision. Therefore, on the one hand, the full admissibility of these instruments responds to the canons of efficiency and cost-effectiveness of the administrative action where the use of an IT procedure that leads directly to the final decision in fact entails numerous advantages such as, for example, the significant reduction in timing procedural for purely repetitive and non-discretionary operations, the exclusion of interference due to negligence (or worse malice) of the official (human being) and the consequent greater guarantee of impartiality of the automated decision. On the other hand, however, the use of computerized procedures cannot be a reason for circumvention of the principles that conform our legal system and that govern the conduct of administrative activity. In this context, in fact, the use of the algorithm must be correctly framed in terms of an organizational module, a procedural and preliminary instrument, subject to the typical checks of every administrative procedure, which remains the modus operandi of the authoritative choice, to be carried out on the basis of legislation conferring the power and the purposes attributed by the same to the public body, holder of the power. In this regard, the Board reiterated that the 'multidisciplinary characterization' of the algorithm (construction which certainly requires not only legal skills, but technical, IT, statistical, administrative) does not exempt the 'technical formula' from the need, which in fact represents the algorithm, is accompanied by explanations that translate it into the 'legal rule' underlying it and that make it legible and understandable. Therefore, the knowledge of the algorithm must be guaranteed in all aspects: from its authors to the procedure used for its processing, to the decision mechanism, including the priorities assigned in the evaluation and decision-making procedure and the data selected as relevant. This is in order to be able to verify that the criteria, conditions and results of the robotic procedure comply with the requirements and purposes established by law or by the administration itself upstream of this procedure and so that the methods and procedures are clear - and consequently can be rules under which it was set.

Conclusions
In light of these premises, it follows that given the general admissibility of these instruments, two pre-eminent aspects assume fundamental importance, also in the light of the discipline of supranational origin, as elements of minimum guarantee for any hypothesis of use of algorithms in decision-making. publishes: a) full knowledge upstream of the module used and the criteria applied; b) the imputability of the decision to the body holding the power, which must be able to carry out the necessary verification of the logic and legitimacy of the choice and the results entrusted to the algorithm. In this context, even in the face of a knowable and understandable algorithm, which is not the only reason for the decision, it must also not assume a discriminatory character.

P.I. 00777910159 – Copyright Il Sole 24 Ore – All rights reserved