Predictive artificial intelligence and the preliminary investigation: A democratic control on the activity of the public prosecutor?
Autor: Mario Mazza
Università degli studi "Federico II" di Napoli
Modalidad: Presencial
Abstract:
The author believes that through modern AI systems it is possible to obtain a prognostic judgment on the preliminary investigation activities of the public prosecutor in a criminal proceeding and to examine whether these actions can actually lead to an indictment and a future conviction against the defendant. It focuses on how AI can be put at the service of the criminal procedure code.
The preliminary investigation phase is characterized by the exponential expansion of the powers of the public prosecutor, the true main subject of this procedural phase, who for the investigative acts can theoretically have at his disposal the enormous force, represented by the judicial police and often with a bulimic resort to telephone interceptions.
Legislative Decree no. 150/2022 introduced new regulations regarding the duration of investigations and their extension, providing for the possibility, upon request of the public prosecutor, of a single extension.
In this phase, the rights of the defense are considerably flattened - perhaps even annihilated - in the face of this exponential expansion of power at the PM's disposal. So much so that with a view to the democratic balance of power, between checks and balances, a body for controlling the PM's activity, called the Judge for Preliminary Investigations, was included in the famous "Vassalli" code.
Unfortunately, the GIP too often is oblivious to the theses of the Prosecutor, both at the time of the preliminary investigations and in the subsequent phase, if foreseen, of the GIP hearing, in which he will decide on the indictment. There is, in truth, a strong relationship between PM and GIP, who come from the same role and frequent the same offices. Furthermore, over time the magistrates have established their own "political" representations such as the “ANM”, “Unicost”, “Democratic Magistracy” and similar with the effect that a PM and a GIP, who manage the same procedure, can belong to the same "political" current, with all due respect to the principle of impartiality.
The eminent constitutionalist Sabino Cassese indicated that, "the prosecutor's offices have acquired a particular place, so much so that many experts speak of a 'Republic of PMs', which have become a separate power, with their own means, which are addressed directly to public opinion , making use of the 'fairy tale' of the mandatory nature of criminal action, using judicial reporting as a means of political struggle and transforming Italy into a 'judicial republic'.”
Only the recent rule on the separation of careers has provided for an effective dividing barrier between the judging and investigative careers: it will be necessary to see to what extent, and when, there will be an effective division of roles, separated from the personal connection.
Well, already in this phase we can feel the useful intervention of artificial intelligence which prepares an elaborate and in-depth analysis of the matter for the GIP, also giving a possible prognostic judgment on the functionality of the extension to the preliminary investigations and for a possible indictment of the 'investigated. In essence, alongside the PM's request for an extension, a report drawn up by the AI could be useful to understand whether the further investigations are actually functional to the procedure and, above all, whether the procedural deadlines have not been exceeded, this last circumstance being much more frequent than you might think.
The author, after having focused on the duration and extension of the deadline for preliminary investigations, examines the pathological effects of investigative activities carried out outside the deadline which give rise to the absolute nullity of the documents and the consequent unusability of the same. He concludes by highlighting that we can therefore perceive the key role that predictive artificial intelligence could play, through a scientific control on the timetable of preliminary investigations, evaluating whether the elements collected were "collected in time" and whether they can presumably and with a prognostic judgment give result in a conviction, with a significant saving in procedural costs and thus deflate criminal proceedings, some stillborn due to the almost total absence of real control in the preliminary investigation phase.
There would be the implementation of the principles referred to in art. 6 CEDU, according to which a trial that exceeds the time reasonably necessary to reconstruct the facts and any responsibilities no longer makes the implementation of the right of defense is profitable or completely impossible, frustrates the exercise of the right to evidence, sterilizes the implementation of cross-examination and, at the same time, allows a possible punitive response to no longer satisfy the constitutionally imposed re-educational purpose, making thus the punishment is no longer socially useful (CEDU 18/3/2021, P. c. Italy, in Dir. Pen. e Processo, 2021, 6, 841; ECHR 7/12/2017, A. c. Italy).
On this point, the countless appeals presented to the European Court of Human Rights by those who had to suffer the unfair and excessive duration of the preliminary investigations would be avoided, with Italy condemned to pay the costs and compensation for damages.